mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-09-01 01:27:50 +02:00
rpc_tensor
10690
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Aman Gupta
|
28bd87b2cd | rpc: allow -sm tensor | ||
|
Aman GuptaandGitHub
|
57291f2644 |
ggml: allow passing alloc dependencies in graph_optimize (#27301)
* ggml: allow passing alloc dependencies in graph_optimize * add alloc dep tests * add TODO about using flat arrayb10689 |
||
|
codemonkeyandGitHub
|
c589f0ed10 | metal : add fa-vec tunings for M2 (#27940) b10688 | ||
|
|
c841aeeb8b |
opencl: use a better matmul path on two Adreno GPU generations (#27640)
* opencl: default the Adreno xmem F16xF32 GEMM on for X2E kernel_mul_mm_f16_f32_l4_lm is the slowest matmul this backend has on Adreno: on the X2-90 it runs the gpt-oss-20b attention projections at roughly a quarter of what the tuned dense q4_0 GEMM reaches on the same device. That matters for any model whose non-expert weights stay f16 -- the stock gpt-oss-20b release is exactly that, and its prefill spends 40.8% of GPU time in that one kernel. The xmem route already existed but was left opt-in, so nobody hit it. Worth about 25% prefill on gpt-oss-20b on an Adreno X2-90. Gated to X2E: the Adreno 840 measures neutral. Decode is untouched -- the dispatch gate needs N >= 16. It is worth nothing on the q8attn variant, whose attention weights already take the dp4a dense GEMM. The env var was presence-tested before, so =0 previously enabled it; it is now atoi()'d. MUL_MAT 963 OK / 0 FAIL on both arms. * opencl: bypass the tiled f32 GEMM on the Adreno A7X The A7X (E031.41) compiler executes kernel_mul_mm_f32_f32_l4_lm at roughly a tenth of what the same silicon reaches in its own f16 and q4_K kernels. It allocates 488 B/WI of private memory against 304 for the same source on the following generation, i.e. the older register allocator spills in the K-loop. Models with per-layer F32 projection pairs kept F32 by quantization policy land on this kernel twice per layer, and it dominates their prefill on that part. Route batched f32xf32 (ne11 > 8) around the tiled path on the A7X and let it fall through to the per-row f32 kernel, which that compiler handles fine; small batches keep the tiled path. Weights stay GPU-resident, so decode placement is untouched -- declining the op in supports_op instead was measured first and rejected, because the per-layer CPU round-trips cost more decode than the prefill it gained. Worth about 9% prefill on gemma-3n-E4B on an Adreno 740, with MUL_MAT counts identical on and off. No other generation is affected. Override with GGML_OPENCL_A7X_F32_LM_BYPASS=0. * opencl: enable xmem GEMM for adreno by default --------- Co-authored-by: Li He <lih@qti.qualcomm.com>b10687 |
||
|
Georgi GerganovandGitHub
|
3173a56471 |
metal : assert shared memory padding (#27951)
* metal : assert shared memory padding * cont : add refb10686 |
||
|
Niklas WenzelandGitHub
|
17252c769a | metal : add remaining fa-vec tunings for M4 Pro (#27915) b10685 | ||
|
Nick FarrellandGitHub
|
cc83d7b482 |
sycl: make --fit respect --fit-target better (#27629)
improve the --fit algorithm to take into account the actual peak required VRAM for a given context size on a SYCL backend. This includes both properly accounting for how much VRAM is required when the allocated context is fully used (which makes the reported context drop below what it did before, but stop it OOMing) as well as preventing some overly-conservative calculations which meant too much VRAM was being reserved. Tested on a Arc b70 with unsloth's qwen3.8 (Q4_K_XL), able to get 262144 context, fully usable, with q8_0 KV and MTP and 4k ubatch size using --fit-target 1b10684 |
||
|
Jeff BolzandGitHub
|
c9ca51c1f6 |
vulkan: combine duplicated fastdiv functions, rename the one optimizing small divs (#27526)
* vulkan: combine duplicated fastdiv functions, rename the one optimizing small divs * remove one more fastdivb10683 |
||
|
Jhen-Jie HongandGitHub
|
5ea1b124e7 | metal : add fa-vec tunings for M1 Max (#27932) b10682 | ||
|
Jeff BolzandGitHub
|
77f132cb1d |
vulkan: Change mul_mat_id to pad K rather than N (#27925)
The N padding is needed for mul_mat, but not mul_mat_id. For mul_mat_id, we indirect the row index through a shared memory lookup table which avoids any OOB row coordinate. But that callback doesn't bounds check K, so we actually need K padding instead.b10681 |
||
|
|
d7bd3bfcad |
snapdragon: python SDK setup (Windows) (#27903)
* port setup-build.ps1 to setup_sdk.py, to facilitate installation of Hexagon and OpenCL SDKs on Windows * rename setup_sdk.py -> setup-sdk.py * flake8 fix: print() -> logger.info() --------- Co-authored-by: Kristopher Urquhart <kurquhar@qti.qualcom.com>b10680 |
||
|
Xuan-Son NguyenandGitHub
|
50f068ffff |
bench: add --tensor-read-lazy (#27881)
* bench: add --tensor-read-lazy * rm the alias * rename to LLAMA_LAZY_MODE_*b10679 |
||
|
Xuan-Son NguyenandGitHub
|
6fe7498016 | model: qwen4exp: reduce number of graph splits (#27880) b10678 | ||
|
|
b387ddfd84 |
vulkan: fix missing view-alias dependencies in ggml_vk_graph_optimize (#27812)
* vulkan: fix missing view-alias dependencies in ggml_vk_graph_optimize is_src_of doesn't treat two views of one tensor as dependent, so the optimizer reorders nodes across aliased reads and writes. Result: silently wrong tokens under greedy decoding, different output on every server start, and invalid speculative-decoding acceptance, with nothing logged. Hits Qwen3.8's recurrent state (and any model with view-aliased state) on AMD and NVIDIA Vulkan. CUDA is clean. Compare view_src bases on both sides. Fixes #27805 * vulkan: don't treat view/no-op nodes as aliasing dependencies Nodes whose op is NONE, RESHAPE, TRANSPOSE, VIEW or PERMUTE execute nothing, so aliasing through them is not a real dependency. The previous base comparison matched them anyway, which only costs the optimizer reordering freedom. Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * vulkan: make the lambda parameter const and capture is_empty in is_src_of Code will not compile without these changes. is_src_of has an empty capture list, so is_empty was not visible inside it, and is_empty took a non-const pointer, while is_src_of receives const ones. Other call sites pass non-const pointers, which still convert as usual. --------- Co-authored-by: Jeff Bolz <jbolz@nvidia.com>b10677 |
||
|
|
a43c3986b4 |
ggml : fix conv_transpose_2d for multiple batches (#26132)
* ggml : fix conv_transpose_2d for multiple batches ggml_compute_forward_conv_transpose_2d_impl only computed the first batch (ne[3] of the destination); every batch after the first was left as zero. Both the src1 permutation and the main compute loop now iterate over the batch dimension, and the work buffer size in ggml_graph_plan is scaled by the src1 batch count so the extra permuted batches fit. A multi-batch test case is added to test-backend-ops. Fixes ggml-org/ggml#1448 * metal : fix conv_transpose_2d for multiple batches The kernel only computed batch 0 of the input (src1->ne[3]); every output batch after the first was left as zero, so multi-batch conv_transpose_2d results diverged from the CPU reference. The grid now covers all batches (OW x OH x OC x N), the kernel decodes the batch from the grid z coordinate and offsets both the input and destination indices accordingly. nb3 is passed in the kernel args. Assisted-by: pi:llama.cpp/Qwen3.8-27B --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>b10676 |
||
|
|
90c26fcd4b |
Vulkan: add hoisting support for row IDs and expert count in shaders (#26686)
* vulkan: add hoisting support for row IDs and expert count in shaders * use hoisted row ids in coopmat2 * vulkan: address review feedback on count_experts - use vk_op_count_experts_push_constants instead of a raw uint vector - apply the fastdiv trick to the ne00 div/mod in count_experts - compute the per-expert offsets with subgroupExclusiveAdd when the device supports it, keeping the serial path as fallback - document the data_d layout and the hoisted_row_id_words bound - drop a leftover debug print in ggml_vk_matmul_id * vulkan: use init_pushconst_fastdiv for count_experts push constants * vulkan: refine comments for row ID hoisting and data layout in count_experts shader * Whitespace --------- Co-authored-by: Jeff Bolz <jbolz@nvidia.com>b10675 |
||
|
Georgi GerganovandGitHub
|
8663224818 | context : disable non-fused GDN and LID ops (#27877) | ||
|
|
f5e85d43a0 |
metal : add fa-vec tunings for M4 (#27875)
This adds fa_vec_tuned_table records for Apple M4 to ggml-metal-tuning.cpp. Includes F16, Q4_0, Q4_1, Q5_0, Q5_1, and Q8_0. (M4, 10 GPU Cores) Co-authored-by: Strongtut <8432058+Strongtut@users.noreply.github.com>b10673 |
||
|
|
511f9c1379 |
OpenVINO: Update OV to 2026.3.1, whisper.cpp support, Qwen3.5 on NPU, and new ops (#27843)
* OpenVINO Backend: Fuse IM2COL + MatMul convolution into OpenVINO convolution * ci:ggml-ov: Skip recurrent state rollback tests * ci:ggml-ov: Skip recurrent state rollback tests * Update OPENVINO.md * ggml-openvino : add env-var gated op support debugging * Fix ggml_rope_set_offset case * OpenVINO backend: Support Whisper.cpp * Fix code style * openvino : enable qwen35 on NPU Static shapes: - get_graph_input_shape() left the s_copy / s_copy-leaf inputs dynamic ([1,1,1,-1]) even in static mode, which propagated a dynamic slot dim through GET_ROWS into the conv/GDN state, the state reshapes and the GDN output. - With -np 1 the s_copy defrag remainder gathers zero rows; short-circuit that CPY to the untouched cache instead of emitting a degenerate Slice/Concat, and skip binding its zero-byte ggml tensor as an output (the dynamic path already did the latter, the static path wrote the full cache over a 0-byte buffer). Token-count independence: - In static mode the compiled model's token count is the prefill chunk size or 1, not the captured cgraph's. Offsets derived from the captured count were therefore wrong. Anchor the GDN state slice at the end of the packed [attn | state] output and drop the rs_src_begin runtime inputs, and make VIEWs over the GDN output / conv_input pass through so the consumer does the slicing. - CONT could not identify its token axis when the graph was captured with a single token (every trailing dim has the same stride and size 1) and baked the captured shape into the prefill model. Chunked prefill: - The last chunk is padded with fabricated tokens. Attention masks them, but the recurrent path folded them into cache_r/cache_s permanently. Add a chunk_valid_len runtime input, use it to zero g and beta for padded steps (making the recurrence an exact identity) and to end the conv snapshot window at the last valid token, and disable the recurrent-cache reset after the first chunk so earlier chunks are not wiped. - get_is_prefill() and the chunk loop bound read inp_pos->ne[0] directly, but IMROPE stacks 4 position planes, so every decode step was run through the padded prefill model and the loop ran extra out-of-bounds chunks. cache_rs_reset_idx/len now stay runtime Parameters in static mode, since can_reuse_statically() does not invalidate the cached model on ComputeParams changes. Add GGML_OPENVINO_FORCE_STATIC to exercise the static path on CPU. * Update to OpenVINO 2026.3.1 * ggml-openvino: forward NPU compilation mode parameters Add GGML_OPENVINO_NPU_COMPILE_CONFIG to the backend's cached environment so callers can configure the NPU compiler without using the generic property escape hatch. When the value is non-empty, pass it to OpenVINO as NPU_COMPILATION_MODE_PARAMS. This enables settings such as optimization-level=3 for NPU compilation while preserving the existing behavior when the variable is unset and leaving CPU and GPU configuration unchanged. Document the variable, its NPU-only scope, and the optimization-level=3 example in the OpenVINO backend runtime configuration table. * ggml-openvino : support RELU, POOL_2D, QUICK_GEGLU, and ROLL ops * reorder op table * exclude GPU/NPU failing POOL_2D case * move op type detection to compute_op_case * Relax rope supported cases * Fix pool case * Update openvino doc, gpu driver in ov docker * openvino: remove unused static remote context branch * openvino: parallelize static model build * Apply editorconfig --------- Co-authored-by: Mostafa Faheem <mostafaaafaheem@gmail.com> Co-authored-by: Ravi Panchumarthy <ravi.panchumarthy@intel.com> Co-authored-by: zhaixuejun1993 <xuejun.zhai@intel.com>b10672 |
||
|
Xuan-Son NguyenandGitHub
|
b19cbe925b | convert: prevent ndarray conversion in LazyChunkedTensor (#27869) | ||
|
Ozymandias_EBONandGitHub
|
d077b4c214 |
sycl: use TILE for quantized KV decode on BMG (#26689)
Route quantized KV decode to TILE on Xe2 (BMG) only, keep VEC on other archs until validated there.b10670 |
||
|
TitaniumtownandGitHub
|
be876204aa |
sycl: bind the f16 KV cache in place for the oneDNN SDPA path (#27468)
Measured at a live KV length of 34816 (32768 depth plus one 2048 ubatch),
on Qwen3.8 27B Q4_K_S:
per tensor 4 * 34816 * 256 * 2 B = 71.3 MB
staged per call K and V, so 2x = 142.6 MB
traffic per call read once, write once = 285.2 MB
traffic per ubatch 285.2 MB * 16 calls = 4.56 GB
One ubatch is one ggml_cgraph submission (llama_context::process_ubatch ->
graph_compute), so that 4.56 GB is the cost of a single 2048-token prefill
chunk, and it scales with the live KV length: the first ubatch of the same run,
at seq = 2048, moves 0.27 GB.
Reproduce the two measured inputs with:
GGML_SCHED_DEBUG=2 llama-bench -m MODEL -p 8 -n 0 -r 1 -ngl 0 \
-fa on -ctk f16 -ctv f16 -v > nd.txt 2>&1
grep -E 'n_layer|n_head_kv|n_embd_head_k' nd.txt
awk '/node # 0 /{g++} g==1 && /\(FLASH_ATTN\)/{n++} END{print n+0}' nd.txt
b10669
|
||
|
Georgi GerganovandGitHub
|
8963a9bdcd |
metal : add fa-vec tunings for M3 Max, M5 and M5 Pro (#27863)
* metal : add fa-vec tunings for M5
This is a followup contribution to
b10668
|
||
|
Brad SmithandGitHub
|
6d6b697cd5 |
metal : add fa-vec tunings for M4 Pro (#27824)
This is a followup contribution to
b10667
|
||
|
Georgi GerganovandGitHub
|
4e97ac86eb |
tests : run test-save-load-state across all architectures (#27755)
* tests : run test-save-load-state across all architectures test-save-load-state previously only ran in ctest against a single downloaded model (tinyllamas/stories15M), i.e. only the llama arch. Add a --models DIR mode to test-save-load-state that runs the full save/load suite over every *.gguf in a directory, reporting a per-model PASS/FAIL and exiting non-zero if any model fails, and wire a ctest to run it over all architectures using the existing generate-models fixture (test-llama-archs). The single-model -m mode is preserved (still used by ci/run.sh). Also bump the dummy-model training context in test-llama-archs from 128 to 256 so that the per-sequence context (which is padded up to a multiple of 256) no longer exceeds n_ctx_train and emits the "possible training context overflow" warning. The test is expected to fail until the affected arches are fixed: deepseek4 (host seq-copy), gemma2/gpt-oss/lfm2 (device seq-copy), minimax-01 (state load). It aborts at the first arch that crashes. Assisted-by: pi:llama.cpp/Qwen3.8-27B * tests : match dummy DSA indexer to fused Lightning Indexer kernel The dummy DSA indexer (deepseek32, glm-dsa, ...) used key_length=64 and head_count=1, so the fused Lightning Indexer op's q tensor was shaped [64, 1, ...]. The Metal fused kernel is fixed to DK=128, NH=64, so it rejected the op and the scheduler fell back to CPU, emitting a 'layer assigned to MTL but Lightning Indexer on CPU' warning. Bump key_length to 128 and the DSA head_count to 64 so the fused op runs on the GPU. Assisted-by: pi:llama.cpp/Qwen3.8-27B * tests : add --help and document -o in test-llama-archs Add a --help/-h flag to test-llama-archs and list the existing -o/--out option in the usage text, which was previously missing. Assisted-by: pi:llama.cpp/Qwen3.8-27B * tests : use 64 indexer heads for deepseek4 deepseek4's indexer head count was set to n_head (8), which does not match the fused Lightning Indexer kernel's fixed NH=64, so the fused op fell back to the CPU backend and emitted a device-mismatch warning. Give it the same fixed 64 as the other indexer archs by dropping it from the n_head ternary (only minimax-m3 keeps n_head, since it does not use the fused Lightning Indexer op). Assisted-by: pi:llama.cpp/Qwen3.8-27B * tests : fix dsv4 save-load n_stream mismatch The dsv4 KV cache keeps per-sequence KV/state streams even in unified mode, so its n_stream equals n_seq_max. The test saved the state in the baseline with n_seq_max=1 but loaded it in the seq-copy tests with n_seq_max=2, so state_read threw an n_stream mismatch. Use n_seq_max=2 in the baseline and state-load tests so the save and load agree. Assisted-by: pi:llama.cpp/Qwen3.8-27B * context : relax on-device seq-copy chunk alignment The on-device state seq copy (llama_state_seq_set_data with LLAMA_STATE_SEQ_FLAGS_ON_DEVICE) copied the write-side cpy tensors to the read-side targets 1:1 by index, requiring the writer and reader to emit the same number of chunks in the same order with the same per-chunk sizes. state_write_data chunks per cell-range while state_read_data chunks contiguous-or-per-cell, so the counts diverged for non-contiguous sources (dsv4, SWA) and the copy aborted with "memory buffer mismatch". All state writers and readers enumerate the same logical data in the same order, differing only in chunking. Copy the flat write-side data into the read-side targets with a byte cursor that walks both tensor lists across their boundaries, so the chunking no longer needs to match. Keep the total-size guard; drop the n_tensors equality check. Assisted-by: pi:llama.cpp/Qwen3.8-27B * model : fix dangling hparams ref in minimax-01 LA graph input llm_graph_input_la stored const llama_hparams & hparams, bound to the llm_graph_params temporary in llama_context::process_ubatch. The input object outlives that temporary (it is kept in llm_graph_result::inputs for graph reuse), so set_input() read destroyed stack memory on every graph reuse - test-save-load-state crashed for minimax-01 when the stack region was overwritten (n_layer_all read as 0, abort in llama_hparams::n_head). Store a copy like every other graph input class. Assisted-by: pi:llama.cpp/Qwen3.8-27B * context : handle "worst case" graph and add TODOb10666 |
||
|
|
ca3d5a3e10 |
model: add DSpark support for Nemotron3.5 (#27804)
* model: add DSpark support for Nemotron3.5 * Update src/models/dflash.cpp Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co> --------- Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co> Co-authored-by: Xuan Son Nguyen <son@huggingface.co>b10665 |
||
|
cqderekandGitHub
|
e70802a01f |
ggml-hexagon: add HTP unary ops for ABS and LOG (#27786)
Add HVX-accelerated implementations for GGML_OP_LOG and GGML_UNARY_OP_ABS on the HTP backend. - Register HTP_OP_UNARY_ABS and HTP_OP_UNARY_LOG in op_remap_to_htp() - Add ABS and LOG to ggml_backend_hexagon_device_supports_op() - Implement hvx_abs_f32_aa() in hvx-arith.h using hvx_vec_abs_f32() - Implement hvx_log_f32_aa() in hvx-log.h using hvx_vec_log_f32() - Add abs_f32() and log_f32() row-wise dispatch in unary-ops.c - Define tiled and non-tiled task functions via DEFINE_UNARY_TASK and DEFINE_UNARY_TILED_TASK macros - Route HTP_OP_UNARY_ABS and HTP_OP_UNARY_LOG through execute_op() in main.cb10664 |
||
|
Aparna M PandGitHub
|
83d855c5a6 | hex-unary: fix RMS_NORM_MUL weight-offset bugs for grouped/broadcast norms (#27798) b10663 | ||
|
|
18443257a3 |
server: add ctx-per-slot (--kv-unified-per-slot) (#24124)
* Add ctx-per-slot argument for unifid KV cache * Swap out ctx fractions for ctx pool slots * Formatting cleanup * Remove ctx-pool-slots, make ctx-per-slot an int * refactor it --------- Co-authored-by: Xuan Son Nguyen <son@huggingface.co>b10662 |
||
|
Harkirat GillandGitHub
|
32176338a6 | ci : build only the ggml-hip backend for windows-rocm release (#27753) b10661 | ||
|
|
6c84c7d5d8 |
model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)
* gguf: add qwen4exp (Qwen3.8-Flash-Next) arch and converter
Adds the GGUF-side plumbing for HF model_type qwen4_exp:
- MODEL_ARCH.QWEN4EXP plus tensors for the low-rank hyper-connection
variant (hc_*_norm/down/up/inject) and the PLE n-gram hash embeddings.
The DeepSeek-V4 hc_*_fn/base/scale tensors are a different
parameterisation, so these are separate entries rather than reuse.
- Reuses the existing indexer, per_layer_token_embd, SSM and
compress_ratios keys unchanged.
- conversion/qwen4exp.py inherits the Qwen3.5 linear-attention V-head
reorder and interleaved mrope, concatenates the 128 PLE embedding
shards, and splits index_qk_proj into separate indexer q/k tensors.
The PLE hash multipliers reach ~2.4e13. prepare_tensors() casts every
non-float dtype to float32 before modify_tensors() runs, and GGUF array
writes infer INT32 from Python ints, so both paths are bypassed: the
constants are read from the pre-cast lazy tensors and written as
explicit UINT64 arrays.
Additive only; no existing arch changes behaviour.
* llama: load qwen4exp (Qwen3.8-Flash-Next) hparams and tensors
Adds LLM_ARCH_QWEN4EXP with its hparams and tensor loading. The graph
comes in the next commit; this makes the model load and report correct
metadata.
- hyper-connections set n_embd_out_impl = hc_count * n_embd, so the
residual stream is 4x wide and there is no output_norm: the final
mixer's hc_norm is the last norm in the model.
- registered as hybrid and given the same recurrent/attention memory
filters as Qwen3-Next and Qwen3.5.
- reuses the existing indexer, per_layer_token_embd, SSM and
compress_ratios keys as-is.
- the PLE table row count is read back from the file rather than
recomputing the vocab padding rule.
llama-model-loader gains UINT64 array support. That branch previously
threw, so no existing caller changes behaviour; it is needed because the
PLE hash multipliers do not fit in int32.
* qwen4exp: shorten comments
* llama: qwen4exp text graph with hyper-connections, GDN and MoE
Implements the decode graph for Qwen3.8-Flash-Next: the hyper-connection
residual stream, gated delta net layers, the MoE block with its gated shared
expert, and dense full attention. The QSA indexer and the PLE n-gram embedding
are not wired up yet and land in later commits.
Hyper-connections are implemented here rather than shared with deepseek4.cpp.
The two formulations agree on the [n_embd, hc, n_tokens] layout and little
else: DeepSeek-V4 mixes with a full-rank projection and Sinkhorn-normalises
it, whereas this model uses a low-rank down/silu/up sigmoid gate and collapses
by a plain mean. Only the ~10 line stream mean is genuinely common, so sharing
would mean touching DSV4's hot path and its three fused CUDA ops to reuse very
little. What is reused is the substantive part: the LLM_KV_HYPER_CONNECTION_*
keys, the n_embd_out_impl wide-residual support already in the loader, and the
layout convention.
Also allows a checkpoint to carry no PLE layers at all, which makes it
possible to bring the graph up and validate it in stages.
Validated against vLLM, the only working reference implementation. On a
scaled-down model with an init scale large enough to give non-uniform logits,
agreement with vLLM sits at the numerical noise floor: llama.cpp f32 against
its own bf16 gives 84.3% top-1 agreement over 255 positions, and this graph
against vLLM gives 85.1%. The comparison was calibrated by seeding three
deliberate bugs (silu instead of sigmoid on the delta net gate, dropping the
1/hc scale in the mix, dropping the 2x in the combine); each drops top-1 to
between 0% and 11%, an order of magnitude below the floor.
* llama: qwen4exp PLE n-gram hash embedding
Adds the per-layer embedding: a custom I32 graph input hashes each token with
its ngram_size-1 predecessors host-side and the result is a plain row gather
over the shared table, the same shape gemma3n's per-layer embedding uses. The
hash has to run on the host because the splitmix64-derived multipliers reach
2^45, so the products need 64-bit integers and an xor, neither of which ggml
has.
Predecessors that fall outside the ubatch come from a small per-sequence
history on the model, mirroring the per-request ngram_context the reference
carries. It is only trusted when contiguous with the incoming position, so a
fresh prompt or a rewound cache falls back to EOS padding rather than hashing
against stale tokens.
The depthwise conv is written out as a sum of shifted, per-channel-scaled
copies rather than through ggml_conv_1d_dw, which carries a correctness
warning upstream.
Verified two ways. The row indices match a transcription of the reference's
tensor formulation exactly, 1024 of 1024 rows, including sequences with EOS
tokens sprinkled through them to exercise the segment reset. Separately, with
PLE placed on layer 0 so its input is just the token embedding, ple_embd and
ple_gated_value match a PyTorch computation from the same checkpoint to every
printed digit.
End to end over 1023 scored positions the port sits the same distance from
vLLM with PLE as without it, 6.3 points of top-1 against 6.0, so PLE costs no
accuracy relative to the rest of the model. That common offset is vLLM's bf16
activations, which cannot be removed: its QSA kernel refuses float32.
Two bugs found along the way, both caught by the row-index check. The history
was read and updated in the same pass, so a token early in a ubatch could pick
up an earlier token of that same ubatch as prior context; it is now snapshotted
first. And an EOS token was cutting its own context, where the reference takes
the last EOS strictly before the position, so a boundary only hides tokens from
the positions after it.
Known gap: the conv carries no state across ubatches, so it is exact only for a
prefill that starts at position 0. Chunked prefill and decode need the conv
state wired into the recurrent memory, and the conv branch itself is still
numerically unverified because the fixture zeroes its weights.
* llama: carry the qwen4exp PLE conv state across ubatches
The PLE depthwise conv was zero-padding on the left, which is only right for a
prefill that starts at position 0. Decode and chunked prefill saw a truncated
history for the first (kernel-1)*ngram_size positions of every ubatch.
The PLE module sits on a layer that is also a delta-net layer, so both need a
conv history in the same recurrent row. Rather than plumb a per-layer state
size through build_rs and build_conv_state, the row is widened once and each
convolution addresses its own slice through a local helper. n_embd_r() gains
the extra span, which is zero for every other architecture because it is
derived from ple_n_heads.
Verified by feeding the same 1024 token sequence in chunks instead of one
shot: at 64 tokens per decode the logits are bit-identical to the single-shot
run, 1023 of 1023 top-1 and a maximum logprob deviation of exactly zero. At
one token per decode they differ slightly, but the no-PLE model differs more
under the same test (94.6% against 97.1%), so that is the usual gemv-versus-
gemm accumulation difference and not the state.
The conv branch is also no longer unverified. With non-zero conv weights the
port sits 6.3 points of top-1 below the numerical floor, the same distance as
with the weights zeroed and as the model with no PLE at all, so the branch
adds no error of its own.
test-llama-archs passes every existing architecture at 0.00e+00, including the
delta-net models that share this code path.
* llama: fix the qwen4exp PLE conv state and unblock test-llama-archs
build_rs writes into the state tensor in place, zeroing one row and copying the
carried-over states, so calling it twice for the same layer let the second call
clobber the first write-back. The PLE layer is also a delta-net layer, so that
is exactly what happened: both convolutions gathered the same row. They now
share a single gather per layer.
The earlier claim that the conv state was carried correctly was tested on a
fixture whose conv weights are zero, where the branch contributes nothing and
chunking matches trivially. Re-running with non-zero conv weights showed the
divergence, growing with the number of ubatch boundaries: 97.1% top-1 at one
boundary down to 90.2% at seven. With the shared gather it is bit-identical to
the single-shot run at every chunk size tried, 512, 128 and 64, with a maximum
logprob deviation of exactly zero over 1023 positions. The delta-net-only model
stays bit-identical too, so nothing regressed there.
Also derive the delta-net conv channel count the way load_arch_tensors sizes
wqkv instead of from ssm_d_inner. The two agree for this model, but n_embd_r()
only bounds the row and the convolution has to match the tensor feeding it.
test-llama-archs previously aborted on this architecture and took every later
architecture with it. qwen4exp is marked MoE-only, given the hyper-connection
keys and an ssm_d_inner consistent with its tensor derivation, and skipped for
now: the hyper-connection keys written by get_gguf_ctx are not reaching the
synthesised file, which needs a separate look. The suite completes again, 124
architectures at 0.00e+00.
* llama: optional indexer key cache in llama_memory_hybrid
Groundwork for qwen4exp's QSA sparse attention. Its indexer needs a per-token
key history for the full-attention layers, but a hybrid model cannot use
llama_kv_cache_dsa: that class derives from llama_memory_i rather than
llama_kv_cache, and llama_memory_hybrid constructs its attention cache
directly. No existing architecture pairs recurrent state with a sparse
indexer, so there was nothing to reuse wholesale.
llama_memory_hybrid therefore gains a third, optional cache, shaped the same
way llama_kv_cache_dsa shapes its lightning-indexer cache: a copy of hparams
with n_head_kv forced to 1 and n_embd_head_k_full set to indexer_head_size.
It is built only when a filter_idx callback is passed, which defaults to
nullptr, so every existing architecture gets exactly what it got before. The
per-sequence operations and the batch preparation forward to it under a null
check, matching how the DSA cache prepares its two caches over the same
ubatches.
test-llama-archs passes all 124 architectures at 0.00e+00, including the 12 in
the hybrid family that share this code. The qwen4exp fixtures are unchanged:
same logits against vLLM, and chunked evaluation still bit-identical to
single-shot.
* llama: QSA sparse attention for qwen4exp
The full-attention layers of this model do not attend to everything. An
indexer scores one mean-pooled key per block of compress_ratio tokens and
keeps a budget of the best blocks, plus the tail of tokens that do not yet
form a complete block. Below indexer_top_k + compress_ratio - 1 cached
tokens every block fits in the budget, so the result is exactly dense.
What is reused rather than rebuilt:
- the mask machinery. build_attn's DSA overload already turns a list of
token indices into a KQ mask via ggml_set_rows, so that block is lifted
out verbatim into build_attn_mask_top_k and shared with a new overload
on llm_graph_input_attn_kv. DSA's node sequence is unchanged; the new
overload exists because llama_kv_cache_dsa assumes MLA and cannot be
dropped into a hybrid model.
- the indexer key cache, which is the optional third cache added to
llama_memory_hybrid in the previous commit. It holds raw keys, because
pooling happens before the norm and the rotation.
The graph expands block scores rather than block indices: giving every
token of a block its block's score needs only a gather, where expanding
indices would need an integer multiply-add that ggml has no op for. Since
the budget is a whole number of blocks and a block's members tie exactly,
the cut still lands on a block boundary.
Everything that depends on cache layout is computed host-side in
set_input_qsa. Blocks are cuts of the position line rather than of the cell
array, so nothing assumes the cache is contiguous.
Measured on the tiny fixture against vLLM, comparing the selected token
indices directly rather than the logits:
below the budget selection identical, and 1024-token logits are
bit-identical to the pre-QSA dense path
above the budget mean jaccard 0.975
The direct index comparison is what made this correct. The reference
rectifies each head's dot product before summing over heads, which an
earlier reading of it had missed; on logits alone the resulting port looked
fine, because on a randomly initialised fixture the known-correct dense
path already disagrees with vLLM by more than the bug did. Comparing the
indices showed 0.794, and fixing the ReLU moved it to 0.975.
* llama: give the qwen4exp indexer cache the attention cache's slots
The indexer cache found its own slots, independently of the attention
cache. Both are the same size and see the same ubatches, so in a
straight-through prefill they agree, which is why every fixture and every
single-shot parity run passed. They drift once the context is being
rewritten between turns, and then the QSA top-k indices, which are applied
against the attention mask, point at the wrong cells.
The seven-turn chat test caught it on the third turn: llama-server aborted
on the assertion that the two caches report the same n_kv.
The cache is a side buffer addressed by the attention cache's cells, so it
now takes that cache's slot layout instead of computing one. Applying that
layout also marks its cells identically, so the two agree cell for cell by
construction rather than by coincidence, and the assertion can no longer
fire.
Inert where the caches already agreed: test-llama-archs green at 126 archs
and 0.00e+00, and the 4096-token tiny fixture is unchanged at max logit
delta 0.0.
* tests: record what the qwen4exp arch-test skip actually observes
The old note guessed that the hyper-connection keys never reach the file.
They do: dumping the gguf_context handed to llama_model_init_from_user
shows both among its 67 KVs, and the loader still reports one missing.
* tests: cover qwen4exp in test-llama-archs
The arch was skipped with a note guessing that the hyper-connection keys
never reached the synthesised file. They did. The suite builds a model, then
saves and reloads it, and llama_model_saver did not re-emit those keys, so
the failure was in the roundtrip leg rather than the first load. Three gaps,
all in shared code and all additive:
- add_kv_from_model wrote no hyper-connection, compress-ratio or PLE keys.
The PLE group only means anything whole, so it is written or omitted
together; the rest follow the file's existing style of writing every key
unconditionally, since an architecture that does not read one is
unaffected by a zero.
- the saver had no uint64 path at all, which the PLE hash constants need.
- add_tensors_from_model enumerates model-level tensors by hand and was
missing per_layer_tok_embd and the three final-mixer tensors.
Two smaller fixes on the qwen4exp side, both found by running the test:
- build_qsa_top_k divided by the compression ratio before asserting it was
non-zero, so a file without the key crashed instead of reporting.
- a layer with no compression ratio now falls back to dense attention,
which is what the model computes below the budget anyway. The test then
has to write a ratio to reach QSA at all, and an indexer key length no
narrower than n_rot, since the indexer ropes with the main attention's
rotary width.
Full suite: 126 archs, qwen4exp at 0.00e+00 with roundtrip OK. The tiny
fixture is unchanged, max logit delta 0.0 against the pre-QSA dense run.
* convert: stream the qwen4exp PLE table instead of concatenating it
The n-gram table arrives as 128 shards that were held in a dict and then
torch.cat-ed, so the peak was the shards plus the concatenation: around
300 GB of RSS on the real checkpoint, which rules out machines that could
otherwise convert this model.
Each shard is now written straight into a memory-mapped file at its final
row offset and dropped, so the resident set is one shard and the rest is
the page cache's problem. The temporary file sits beside the output and is
removed once the write finishes, including on failure.
Shards other than the last must be uniform for direct placement, which is
asserted rather than assumed, and a shard arriving before the stride is
known is held instead of misplaced.
Verified on the tiny fixture: the resulting GGUF is byte-identical to the
one the concatenating path produced (md5 2d274efac91ad1e9a6007efb0687e597).
* quantize: fall back to F16 for 32-block types with an odd ncols
tensor_type_fallback demotes a tensor whose ncols is not a multiple of the
target's block size, but its switch only enumerates the 256-block types. A
target that is already a 32-block type (iq4_nl, q4_0, q5_0, q8_0, ...) falls
into default: and throws, even though the function already knows how to answer
that case: the ncols check right below the switch resolves an unrepresentable
shape to F16.
Route those types into that check instead of throwing. Only paths that abort
today change, so no quantization that currently succeeds is affected.
Found on a 4-wide depthwise conv kernel. llama-quantize reported nothing but
"failed to quantize model from ...", with no tensor name and no exception text,
which made a quant recipe that had simply not pinned the tensor look like a
corrupt model. It now names the tensor and continues.
* quantize: let --tensor-type name per_layer_token_embd
per_layer_token_embd shares the TOKEN_EMBD category with token_embd.weight, so
--token-embedding-type is returned for it before any --tensor-type pattern is
consulted, and there is no way to give it a tier of its own.
That grouping is fine as a default and stays the default. It is a poor fit for
the size, though: on qwen4exp the table is 97.7 GiB of a 337.6 GiB BF16 file and
about 46% of a 4-bit one, roughly eighty times token_embd.weight, and it is
read by ggml_get_rows rather than a matmul so no imatrix ever covers it.
Allow an explicit --tensor-type pattern to name it, and only it. Nothing
changes unless such a pattern is passed, and token_embd.weight keeps the old
precedence in either case.
Measured on Qwen3.8-Flash-Next, Q4_K_M with an imatrix: the table lands at q8_0
(51.9 GiB, 113.5 GiB total) by following --token-embedding-type, and pinning it
q4_1 gives 30.5 GiB for 92.1 GiB total, 19% off the file.
* quantize: size the output buffer exactly instead of nelements * 4
The per-tensor output buffer was sized `nelements * 4`, described as an upper
bound. It is a very loose one: the output is at most 2 bytes per element
(f16/bf16) and usually well under 1.1 (q8_0 and below), so between 2x and 4x of
it is never touched. The exact size is already known here, since it is what the
quantization loop writes, what new_size sums to, and what the GGUF metadata is
asserted against a few lines later.
On a model whose largest tensor is a few GB none of this matters. On
Qwen3.8-Flash-Next it does: per_layer_token_embd is 51.2 G elements, so the
buffer was 205 GB where 54 GB is needed at q8_0 and 32 GB at q4_1.
Measured on that model, VmHWM of a live llama-quantize was 485 GB per process.
Three of them fit in 2 TB and five did not, which is what an OOM-killed quant
ladder looks like. This removes about 150 GB of that.
Byte-identical output, verified against the same binary built at the parent
commit: q4_K, q8_0, q5_K, q6_K and IQ4_XS, over BF16 and F32 sources, with and
without a PLE table present. Six cases, six matching md5s.
* qwen4exp: hash the image placeholder for multimodal batches
The PLE row indices are computed host-side from ubatch->token, and set_input
returned early when that was null. A multimodal ubatch is exactly that case:
the mtmd layer consumes the image placeholder ids and hands llama_decode
embeddings instead. The early return left the I32 index tensor uninitialised,
so ggml_get_rows indexed a 320 M row table with whatever the buffer happened to
contain, and aborted:
GGML_ASSERT(i01 >= 0 && i01 < ne01) failed
ggml_compute_forward_get_rows
mtmd_helper_decode_image_chunk -> llama_decode
Every image request crashed. Nothing caught it because the vision work had only
ever been verified by converting an mmproj, never by running one.
The reference computes the hash over input_ids, where those positions still
hold the image placeholder, so carry that id through as qwen4exp.ple.image_token_id
and hash it. The key is optional: a file converted before it existed falls back
to the PLE EOS token, which is defined and treats the image as a segment
boundary rather than crashing.
Verified end to end with llama-mtmd-cli, a Q4_K_M base and the F16 mmproj, on a
generated image with known content. The model names the red circle, the blue
square, the inverted green triangle and reads "UNSLOTH 42", each with the right
position.
* qwen4exp: support a non-unified KV cache in QSA
set_input_qsa asserted n_stream == 1, so llama-server could not serve this
model with more than one slot unless -kvu was passed. With a non-unified
cache each sequence owns its own cells, and a cell index means a different
token in each stream, so a single shared mapping is wrong.
- cell_blk, blk_cells and bias gain a stream dimension. At n_stream == 1
these collapse to the shapes they had, so the unified path is unchanged.
- Scoring is now batched over streams. ggml_mul_mat matches ne[2] on both
operands, so stream s's queries only ever meet stream s's blocks; without
this sequences would score against each other's context.
- set_input_qsa loops per stream and resolves cells through
v_cells[seq_to_stream[seq_id]], following set_input_kq_mask_impl, instead
of hardcoding v_cells[0].
- llama_kv_cache_context::get_n_stream() is added, mirroring the ns that
get_k and get_v already derive from the slot info.
build_attn_mask_top_k needed no change: it already expects
[n_top_k, n_batch, 1, n_stream], so the top-k result is reshaped to meet it.
set_input_qsa has exactly one caller, so the blast radius is qwen4exp only.
Validation, UD-Q4_K_XL on one B200:
- unified cache unchanged within noise: 1802.9/68.85 -> 1807.2/69.11 t/s at
batch 1, 2262.5/192.43 -> 2270.1/193.75 at batch 4.
- non-unified now runs at npl 1, 4, 16 where it previously aborted, and is
22% faster than the -kvu workaround at batch 16 (1205 vs 984 t/s total),
since per-stream cells avoid the cross-sequence masking a unified cache
pays for.
- no cross-stream contamination: four concurrent sequences each carrying a
distinct secret all recall their own and no other, on both cache modes.
- test-llama-archs green on qwen4exp, deepseek2, gemma3n, qwen3next, llama.
Note on testing: comparing concurrent output against solo output exactly is
not a valid check. It failed 0/4 with no bug present, and the unified-cache
control failed the same way, because batch composition changes the
floating-point reduction order and near-tied tokens flip. The contamination
test above is what the exit code gates on.
* llama: keep the qwen4exp top-k attention mask arch-local
The QSA graph needed a build_attn that attends only to the cells named by a
top_k tensor, and the first version got it by adding a llm_graph_input_attn_kv
overload to llm_graph_context and factoring the mask construction out of the
existing MLA sparse path into a shared build_attn_mask_top_k.
That put a new arch on the shared attention path and made the deepseek32 and
glm-dsa attention build depend on a helper introduced for qwen4exp. Build the
mask in src/models/qwen4exp.cpp instead and leave llama-graph.{h,cpp} exactly as
they were: the MLA path keeps its own copy of the same node sequence.
The nodes emitted are unchanged, so this is bit-identical.
* llama: hold the qwen4exp indexer cache in a new llama_memory_hybrid_idx
The indexer key cache was added by extending llama_memory_hybrid with an
optional third cache, and the host-side cell/block mapping that drives QSA was
added as set_input_qsa on llama_kv_cache. Both are shared classes that every
hybrid and every attention model goes through.
Move both into a new memory type, llama_memory_hybrid_idx, following
llama_kv_cache_msa: the indexer cache and the pos<->cell translation live with
the sparse-attention memory rather than in the classes that serve every other
architecture. llama-kv-cache.{h,cpp} and llama-memory-hybrid.{h,cpp} are
restored to their unmodified state.
init_batch is repeated from llama_memory_hybrid because the indexer cache has to
be handed the attention cache's slot infos, and those are not reachable through
the context the base returns. Allocating them separately lets the two caches
drift, which is what pointed QSA's top-k at the wrong cells before.
The context derives from llama_memory_hybrid_context so build_inp_mem_hybrid
keeps working unchanged, and get_n_stream is computed from the slot infos
exactly as llama_kv_cache_context did.
Behaviour is unchanged: logits over an 8192-token sequence are bit-identical to
the previous implementation, sparse and dense alike.
* llama: save and restore the qwen4exp indexer KV cache
llama_memory_hybrid_idx forwarded clear, seq_rm, seq_cp, seq_keep, seq_add and
seq_div to the indexer cache but not state_write / state_read, so a saved
session dropped the indexer keys and a restored one selected QSA top-k against
an empty cache. The effect is invisible until the context passes
indexer_top_k + compress_ratio - 1 cells, because QSA is exactly dense below
that and the indexer contents cannot change the result.
The indexer section is written last rather than next to the attention cache it
mirrors. As a suffix, a reader that does not expect it stops early and the
trailing bytes are caught by the size check in state_load_file; placed between
the attention and recurrent sections it would instead be parsed as recurrent
state, which can succeed and restore silent garbage. It follows the same
LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY gate as the attention cache, since a partial
checkpoint deliberately skips the token-level attention caches.
The indexer restores its own cells instead of taking the attention cache's
restored slots. The two caches share size, padding and every sequence
operation, and init_batch hands the indexer the attention cache's slot infos,
so both state_read_meta calls run find_slot over identical occupancy and land
on identical cells.
The overrides live on llama_memory_hybrid_idx, the only memory type that owns
an indexer cache, so llama_memory_hybrid and every architecture that uses it
write and read exactly the bytes they did before.
The session and sequence state versions are bumped because the qwen4exp state
layout changed. The session path already rejects a short read via its size
check, but llama_state_seq_load_file accepts one silently, so only the version
check stops a pre-fix blob from being half-restored by a fixed build.
(cherry picked from commit 2721542354f8e158c3217625f4e2e7b83e51e3fe)
* llama: make the qwen4exp PLE n-gram history per context and serialise it
The PLE hash of a token mixes in the ple_ngram_size - 1 tokens before it, which
a decode ubatch does not carry, so they were remembered in a map on
llama_model_qwen4exp. That is the wrong owner twice over.
A llama_model is shared by every context that loads it, and the map was keyed
only by llama_seq_id, so two contexts running the same sequence id - two server
instances on one model, or a draft/target pair - overwrote each other's window.
The next_pos guard turned that into EOS padding instead of a crash, so it
degraded quality silently.
The map was also in no state blob: grep found ple_hist in neither
llama-kv-cache.cpp nor llama-memory-*.cpp nor llama-context.cpp. A restored
context therefore failed the next_pos check on its first ubatch and hashed the
first tokens after the restore against EOS padding. This is why a session blob
round-tripped byte for byte while the restored context computed different
logits: the state was never in the bytes.
It moves to llama_memory_hybrid_idx, which is per context, is the memory type
qwen4exp always builds, and already does the per-sequence bookkeeping this
needs. Every sequence operation now carries the window with it:
seq_rm a rewind (p1 < 0) truncates the window to the surviving prefix and
moves next_pos to p0, so a rollback keeps exact context; a hole
punched in the middle leaves the window non-contiguous, so it is
dropped
seq_cp the destination inherits the source's window, truncated to the
copied position range - a copied sequence continues with the same
n-grams the source would have used
seq_keep every other sequence's window is dropped, like its cells
seq_add a shift that moves the whole window keeps it and moves next_pos with
it, which is the context-shift case; one that cuts through it drops
it
seq_div positions stop being consecutive, so an overlapping window is
dropped
clear everything is dropped
Dropping means next_pos = -1, which set_input turns into full EOS padding: the
same thing a fresh sequence gets, and the same thing this code did before it
followed the sequence operations at all, so no case is worse than before.
The state payload is a self-delimiting list, u32 count then per entry
{ i32 seq_id, i32 next_pos, u32 n_toks, i32 toks[n_toks] }, so a whole-context
save and a single-sequence save share one format and a single-sequence restore
can retarget the window at its destination seq_id. It is written after the
indexer section, last, for the same reason that one is: as a pure suffix an
older reader stops early instead of parsing these bytes as something else.
Unlike the indexer section it is not under LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY.
The window is recurrent state - it is the input the PLE convolution's own
recurrent state is derived from - and the recurrent cache beside it is written
for partial checkpoints too. Gating it would leave the server's speculative
decoding checkpoints restoring the conv state without the window that produced
it.
No further version bump: LLAMA_SESSION_VERSION 10 and LLAMA_STATE_SEQ_VERSION 3
were introduced for the indexer section in the same unreleased series, and both
changes are qwen4exp-only additions to the same blob layout.
Also fixes the padding of a short window. set_input pads a window shorter than
ngram_size - 1 up to that length, but prev() indexes the snapshot with the most
recent token last, and resize() pads at the back, so the filler EOS landed where
the immediately preceding token belongs. It now pads at the front. A window is
short at a sequence start after a one-token prefill, and after a seq_rm rewind,
which the new bookkeeping makes common.
Every architecture other than qwen4exp builds llama_memory_hybrid rather than
llama_memory_hybrid_idx, has no PLE table and never asks for a history, so
nothing about its graph, its sequence operations or its state bytes changes.
(cherry picked from commit de170364c052c68fcf63285cc0028095edb9f23c)
* qwen4exp: tidy comments and simplify image token read
Rewrite the comments this series adds to the AGENTS.md rules: one or two lines,
no prose hard-wrapped mid-sentence, no narrative or history, and no comment that
only restates the code. Net 146 fewer comment lines, no code change.
Correct the PLE image comment: mtmd does not consume the placeholder ids. An
image is decoded as an embeddings-only batch, so ubatch->token is null and the
per-position ids never exist here. gemma3n and gemma4 hit the same case and
stand in row 0 of per_layer_token_embd; qwen4exp stands in the configured image
token id instead.
Read image_token_id straight from self.hparams in the converter. base.py merges
text_config into the root of hparams, and the key sits at the root of
config.json, so the config.json re-read was redundant.
(cherry picked from commit 205840c12169057da3e8d2f65ec4ceec3e18b980)
* qwen4exp: support a quantized KV cache in the QSA attention path
(cherry picked from commit 4c30574f81dc1115d08078c47b6cf8c789c0a842)
* llama: give qwen4exp a large-graph node budget
(cherry picked from commit 37c8c194e6a30e4c46ac29bee3fb264f091596ef)
* qwen4exp: drop an unused variable that breaks -Werror builds
(cherry picked from commit 528d032b51fa3cf935ed3ef6e0fb1c7401df53b5)
* quantize: dequantize and quantize large tensors in row bands
f32_conv_buf held the whole dequantized tensor, which is 204.8 GB for
per_layer_token_embd alone and dies with std::bad_alloc long before the
work buffer is reached. Dequantize and quantize in bands of whole rows
instead, capping the f32 staging at 1 GiB per band.
Rows are independent and the imatrix is indexed by column, so band
boundaries cannot change any output byte. Bands nest inside the existing
per-expert loop so each expert slice keeps its own imatrix, and a band is
kept to at least one quantization chunk per worker thread so the existing
multithreading still has work. F32 sources still stage nothing and are
banded by pointer arithmetic into the tensor.
llama_tensor_dequantize_impl now takes a first element offset; the single
caller is updated.
(cherry picked from commit 658c22549613555dbce57a772be4de8509eba3ee)
* llama: segment the qwen4exp fused QKV for tensor split
qwen4exp was missing from the gated delta net branch of get_split_segments,
so its attn_qkv.weight, shaped {n_embd, 2*key_dim + value_dim}, fell through
to the generic fused QKV rule and tripped
GGML_ASSERT(tensor->ne[axis] == n_embd + 2*n_embd_gqa) while loading with
--split-mode tensor. --split-mode layer was unaffected.
qwen4exp broadcasts K to the V heads by tiling, k_conv is grown with a plain
ggml_repeat_4d over the head axis so that v head j pairs with k head
j % n_k_heads. That is the Qwen 3.5 pattern, not the repeat interleave that
Qwen 3 Next builds explicitly, so qwen4exp takes the else branch and its V is
segmented on the scale of K.
Reported by benklop.
(cherry picked from commit 353d753f595dc81634ae6130188b31f06018f5ae)
* llama: fix the qwen4exp PLE history seq_rm(-1) iterator invalidation and the fatal-warning build
ple_hist_rm recursed over ple_hist with a range-based for and the recursive call
erases the entry it is iterating when the whole sequence is removed (p0 <= 0,
p1 < 0), so the loop then increments an invalidated iterator. It is unreachable
today only because llama_memory_recurrent::seq_rm rejects seq_id < 0 before
llama_memory_hybrid_idx::seq_rm reaches the history, which is a guard in another
class. Advance past the entry before recursing.
Two smaller things in the same area:
- the n_toks sanity bound in ple_hist_state_read was the literal 64, which is
the value of LLAMA_MAX_PLE_HEADS, not of the quantity being checked. The
window is at most ple_ngram_size - 1 tokens, so the bound is
LLAMA_MAX_PLE_NGRAM - 1, eight times tighter.
- build_conv_state_at left mem_size unused, so -DLLAMA_FATAL_WARNINGS=ON does
not compile. Predates this series; drop the line.
(cherry picked from commit 6eba44a89d5f328eb4859b844e1d28fb564cbe3e)
* qwen4exp: include llama-impl.h explicitly for llama_mul_mat_hadamard
(cherry picked from commit b634fd4d250d181ef82bf78bd00c1ae3b96a7af6)
* convert: fix the qwen4exp lint and type-check failures
flake8 flagged an unused MmprojModel import, and ty flagged seven errors in
the PLE streaming path: eos_token_id can be absent, and _ple_map, _ple_path,
_ple_row_dim and _ple_rows_per_shard are all Optional at the declaration but
were dereferenced without narrowing.
The map is opened and the stride fixed before the first shard is written, and
_finish_ple_table only runs once every shard has landed, so the invariants
hold. Assert them so the checker can see it. A missing eos_token_id now raises
with the reason instead of a TypeError from int(None).
* llama: give the qwen4exp indexer cache its own tensor names
The indexer KV cache and the attention KV cache both named their tensors
cache_k_l%d, so the Meta backend matched the indexer cache against the
attention split pattern and aborted in handle_set_rows. Tag the names
instead, and mirror the indexer cache: it has one key head and its
projections are mirrored.
(cherry picked from commit a1cdc8181134659766763a17762545a1f0e5db7b)
* qwen4exp: double the Q split granularity for tensor parallelism
qwen4exp fuses the attention gate into attn_q.weight the same way qwen3next
and qwen 3.5 do, so a device boundary must fall on a whole q+gate pair or the
Q heads stop lining up with the K/V heads and attn_output rows.
(cherry picked from commit 6c9a592f0a425a459ab6efae3b897cf68460e244)
* qwen4exp: keep the indexer cache in step across server slots
The QSA indexer keeps a side cache addressed by the cells of the attention
cache, so cell j has to hold the same token in both: the top-k indices it
produces are applied to the attention KQ mask. init_batch already hands the
indexer the attention cache's slot layout rather than letting it look for its
own, but the restore path did not. state_read called llama_kv_cache::state_read
on the two caches in turn and each ran its own find_slot over its own occupancy.
That agrees only for as long as nothing has already pushed the two caches apart,
which is the property a restore is supposed to re-establish rather than one it
can lean on.
The failure path was the worse half, and it is reachable from the public API
with nothing more than a short buffer. Truncating a good blob at 35 offsets and
feeding it to llama_state_seq_set_data left the two caches disagreeing at 5 of
them, and every one of 23 truncations of a whole-context blob did. Four of those
five land inside the attention section, so the attention cache drops the
sequence and the indexer keeps it; only the cut that lands in the indexer
section gives the opposite direction. llama_kv_cache::state_read cleans up its
own cache and rethrows, so whichever way it falls, nothing is left to bring the
two back together. The server papers over this by clearing the slot when a
prompt cache load fails; a caller of llama_state_seq_set_data that does not is
left with an indexer addressing cells that no longer mean what it thinks.
llama_kv_cache::state_read_sinfo reports the cells a restore landed in, or takes
a copy of them, and state_read_meta uses a supplied layout in place of find_slot
once it has checked that those cells are free here too. The indexer now adopts
the attention cache's restored layout by construction instead of reproducing it
by coincidence, and a layout that does not fit fails the read rather than being
applied over cells that already drifted. The hybrid restore is wrapped so that
any failure drops the sequence, or for a whole-context restore the context, from
all three caches at once, which is a state they do agree on.
* kv-cache: clear the cache once when restoring a whole context
state_read walks the streams of the cache in turn, and for a whole-context restore
each stream went through state_read_meta, which starts by calling clear(). clear()
resets every stream at once, so each stream after the first threw away the streams
already restored, and the K/V buffers with them. A non-unified cache holds one
stream per sequence, so a context saved with N sequences in it came back with only
the sequence in the last stream that carried any cells - the highest sequence id.
A unified cache has one stream and never showed it.
The cache is now emptied once, before the loop, which is what a whole-context
restore means. A blob whose streams are all empty now empties the cache as well,
where before it left the old contents in place.
* kv-cache: check the mirrored slot layout on a whole-context restore too
state_read_meta only looked at the layout it was given on the single-sequence path.
A whole-context restore lays the cells out from 0 in both caches, so they agree as
long as they restore the same number of cells, but nothing checked that they did: an
indexer section belonging to some other context was read over cells the attention
cache had filled from a different one, which is the state the indexer must never be
left in.
* qwen4exp: give the PLE conv history its own mirrored recurrent row
n_embd_r() reserved n_conv + ple_conv_state() so that one cache_r_l row could
carry both the delta-net conv state and the PLE dilated conv history, but the
QWEN4EXP arm of get_split_segments only described n_conv. Under -sm tensor the
segment sum came up short by ple_conv_state() and llama_memory_recurrent
construction aborted in ggml_backend_meta_alloc_ctx_tensors_from_buft.
Widening the segment list is not the fix. The Meta backend propagates a view's
split descriptor from its parent unchanged, so a view of one sub-range of a
split axis is sized as the whole row on every device; declaring the PLE tail as
a second segment merely moves the abort to "shape mismatch for VIEW" at graph
allocation. The two histories also want opposite policies: the delta-net state
is split by head to match wqkv and ssm_conv1d, while per_layer_tok_embd,
ple_conv1d and ple_norm_conv are all mirrored, so every device computes the
whole dilated conv and needs the whole history. One tensor cannot be both, and
the split state has no per-segment mirroring.
Move the PLE history into its own cache_ple_r_l%d row, mark it MIRRORED, and
return n_embd_r() to n_conv. The row is allocated only on layers where is_ple
holds, so mirroring one 92160-element row per device replaces a 92160-element
tail on all 36 recurrent rows: the recurrent R footprint drops rather than
grows. build_conv_state_at now takes its width from the tensor it was handed
and keys its gather on that tensor, which also drops a cont of a strided view.
* no more ple_hist (use master version)
* llama: give the qwen4exp full memory context its indexer cache
graph_reserve() walks a full memory context, and qwen4exp builds its
sparse attention only when the context exposes an indexer cache. the
full-context constructor left ctx_idx null, so the reserved worst case
was the dense fallback: a smaller graph than the one decode executes.
ggml-alloc then had to grow the compute buffer on the first decode,
past the size reported at load.
with -np 4 -c 32768 -fa on -ctk q8_0 -ctv q8_0 on an IQ1_S qwen4exp,
the reserved CUDA0 buffer was 217.00 MiB against 275.71 MiB actually
used, and CUDA_Host 42.31 MiB against 191.14 MiB. reserving the sparse
graph makes both match exactly, in unified and non-unified cache mode.
Co-authored-by: Pascal <admin@serveurperso.com>
Assisted-by: Claude
* qwen4exp: shrink the PLE hparams storage
llama_hparams is held by value inside llm_graph_params and every llm_graph_input_*,
and llm_graph_params is a stack local in graph_reserve and process_ubatch, so its
width is paid on every worker thread stack.
is_ple_impl spent 2048 bytes carrying 512 bits. It is the one per-layer flag that is
not moved through the loader's uint32 array templates, so a bitset costs nothing in
call sites and also removes the uninitialized read that non-qwen4exp archs had, since
nothing filled the array for them.
The PLE head offsets and vocab sizes are token-space indices; the gather that consumes
them already truncates to int32, so 64-bit storage was never reachable. The gguf arrays
stay uint64 for file compatibility and are narrowed on load.
sizeof(llama_hparams) 34440 -> 31944, sizeof(llm_graph_params) 34872 -> 32376.
* llama: opt-in random-access mmap advice for host-resident gather tables
qwen4exp keeps per_layer_token_embd on the host: 26.8 GiB at IQ4_NL, read
by ggml_get_rows as 16 gathers of ~90-170 bytes per token, spread across
16 head regions ~20M rows apart. Measured over 4.75M gathers, no two
consecutive gathers land on the same 4 KiB page, so the readahead the
loader asks for buys nothing here and the whole table ends up cached to
serve about 4% of itself.
llama_mmap applies POSIX_FADV_SEQUENTIAL, MAP_POPULATE and a whole-file
POSIX_MADV_WILLNEED unconditionally. Those are right for streaming the
file once into buffers and wrong for whatever stays mapped afterwards.
Under LLAMA_MMAP_RANDOM the eager pull-in is skipped and the mapping is
advised random once every tensor has been read, so the load itself keeps
its sequential readahead. That alone drops the table to 4.4% resident but
serializes one NVMe latency per gather.
The second half is what pays for it: the PLE input already computes every
row index for the ubatch before the graph runs, so the pages those rows
fall on are handed to the kernel in one batch and the reads overlap.
POSIX_MADV_WILLNEED on POSIX, PrefetchVirtualMemory on Windows, which
takes the discontiguous ranges in a single call.
Off by default and off for every other model: the batched prefetch keys
off "this mapping was advised random", which nothing sets unless the user
opts in.
-c 512 --chunks 60, cold, IQ1_S, mean of 3:
default 35.3 s 26.82 GiB resident (100%)
advice only 104.5 s 1.19 GiB resident (4.4%)
advice + prefetch 34.2 s 1.19 GiB resident (4.4%)
PPL 4.2346 +/- 0.07862 in all three. IQ1_S KLD is unchanged in every
field, including Mean KLD 0.396070 +/- 0.001931 and Same top p 77.325%.
* llama: narrow the random-access mmap advice to the gather table
The advice was applied per mapping: every mapping the model kept got
POSIX_MADV_RANDOM plus a whole-file POSIX_FADV_RANDOM, and the eager
pull-in was skipped for every file. On qwen4exp that also hit
token_embd.weight, which sits 0.33 GiB past the PLE table in the same
shard and is read densely, not by sparse gathers. Measured over
-c 512 --chunks 60 on IQ1_S it fell to 8.45% resident, against 100% with
the feature off.
A model now nominates its gather tables (qwen4exp: per_layer_tok_embd)
and only those byte ranges are advised. The range is rounded out to
whole pages, which on this model takes in 832 bytes before and 192
after. token_embd goes back to 86.55% resident and the PLE table still
drops to 4.44%; smaps shows one VM_RAND_READ VMA of exactly the table
instead of one over all 27.16 GiB that stays mapped.
posix_fadvise is dropped from the narrowed path. POSIX_FADV_RANDOM
ignores its offset and length and marks the whole open file, and the
FMODE_RANDOM it sets is only read by page_cache_sync_ra() on the read()
path, which a fault on a MADV_RANDOM vma never reaches. POSIX_FADV_
DONTNEED does take a range, so the drop mode keeps it.
The eager pull-in is now skipped only for the files holding a nominated
table, and re-issued as WILLNEED over the rest of such a file, so other
shards load exactly as before.
prefetch_rows() keys off the tensor being nominated rather than off a
mapping-level flag, so the batched readahead lands only where the advice
did.
-c 512 --chunks 60, cold, IQ1_S, mean of 3, total wall:
default 32.50 s
whole mapping 30.05 s
narrowed 30.35 s
PPL 4.2061 in all three. IQ1_S KLD is bit-identical with the feature on
and off, including Mean KLD 0.396070 +/- 0.001931 and Same top p
77.325%. tg128 73.65 +/- 0.33 narrowed against 73.49 +/- 0.34 whole.
Assisted-by: Claude
* llama: fold the random-access prefetch into its own feature flag
LLAMA_MMAP_RANDOM_PREFETCH existed to measure the two halves of the feature
apart, and the measurement is done: on a cold cache over the same wikitext
run, MADV_RANDOM without the batched readahead takes 94.4 s against 36.7 s
for an untouched mapping, while the pair together take 34.1 s. Suppressing
the kernel's readahead only pays if we replace it, so the split let a user
select a 2.6x regression through a documented switch.
Keep the accessor, since the call site reads better than a mode comparison,
but derive it from the mode alone.
* FACP (Fewer Acronym Classes Please)
* qwen4exp: bias the QSA selection per block, not per cell
The QSA bias is a graph input, so it is pinned on the host and uploaded every
decode, and at -c 32768 -np 4 its twelve copies were 768 of the 815 MiB of
reserved host compute buffer.
Only one half of it needs a cell: whether the cell sits in the always-visible
tail, and whether its block was pooled. Both are properties of the block. The
other half - empty, other sequence, or in the future - is the plain visible/not
test the attention mask already carries over the same cells, so add that mask
instead of repeating it. The bias then holds one value per block.
A block sits wholly inside or wholly outside the tail because the tail starts on
a block boundary, so one value per block is exact. Cells no block covers keep
their -inf from the mask.
The mask is F16 and the bias F32, and a mixed ggml_add reinterprets the F16
buffer as float rather than converting it, so the cast is required.
reserved host compute buffer at -c 32768 -np 4:
--kv-unified 814.86 -> 238.86 MiB, CUDA0 721.07 -> 421.07 MiB
--no-kv-unified 214.86 -> 70.86 MiB, CUDA0 317.07 -> 265.07 MiB
Selection is unchanged: over 8192 tokens, four times the budget, every QSA
layer returns identical top-k indices and the logprobs are bitwise equal.
Two things a reviewer should know. A cell whose position divides past the last
block is guarded by an assert rather than handled, because no run reached it.
And the mask's same-position M-RoPE rule cannot fire for text and was never
exercised for images, so the 2D case is unverified.
* clean up code comments
* clean up new comments
* revert LLAMA_MMAP_RANDOM
* nits
* replace some changes with #27795
* improve the m-rope image for get_prev_tokens
* LazyChunkedTensor
* fix lint
* add some validations
* reduce input nodes
* trim output tokens
* nits
* some more sanity checks
* fix llm_graph_input_ple reuse
* exclude from webgpu test
---------
Co-authored-by: danielhanchen <danielhanchen@users.noreply.github.com>
Co-authored-by: danielhanchen <unslothshared@gmail.com>
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
Co-authored-by: Pascal <admin@serveurperso.com>
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
b10660
|
||
|
Slobodan JosicandGitHub
|
6fdd0ac890 |
ci : bundle HIP runtime DLLs with Windows ROCm release (#26973)
Copy amdhip64_7, amd_comgr and rocm_kpack next to the binaries so the correct HIP runtime loads over the driver's copy in System32. Fixes #26929.b10659 |
||
|
|
b10f9ca58c |
spec : add DFlash2 support (local convolution + candidate selector) (#27342) (#27816)
* spec : add DFlash2 support (local convolution + candidate selector) (#27342) * support DFlash2 * Add p_min in DFlash2 Assisted-by: Claude Opus 5 * Revert unnecessary changes Assisted-by: Claude Opus 5 * Revert draft sampling in rejection sampling Assisted-by: Claude Opus 5 * Refactor code structure Assisted-by: Claude Opus 5 * Delete embedding scaling Assisted-by: Claude Opus 5 * Gate output transforms on DFlash2 Assisted-by: Claude Opus 5 * Optimize Dflash 2 cost Assisted-by: Claude Opus 5 * Avoid using atoi Assisted-by: Claude Opus 5 * Modify comments Assisted-by: Claude Opus 5 * Move llama_model_dflash_selector_top_k to llama-ext.h Assisted-by: Claude Opus 5 * Formatting Assisted-by: Claude Opus 5 * Apply patch to fix the mrope bug Assisted-by: Claude Opus 5 * fix ci Assisted-by: Claude Opus 5 * Fix graph number calculation Assisted-by: Claude Opus 5 * rename hid and unary Assisted-by: Claude Opus 5 --------- Co-authored-by: Jian Chen <jianchen0311@gmail.com> Co-authored-by: Xuan-Son Nguyen <son@huggingface.co> * revert top-k.cu changes --------- Co-authored-by: Zihan Zhang <tiancaizhangdaxian@sjtu.edu.cn> Co-authored-by: Jian Chen <jianchen0311@gmail.com>b10658 |
||
|
Shawn GuandGitHub
|
58546250cf |
opencl: add bin kernels kernel_gemm_moe_q4_0_q8_1_dp4a_bin, kernel_gemm_moe_mxfp4_q8_1_dp4a_bin (#27768)
b10657
|
||
|
Xuan-Son NguyenandGitHub
|
732707dff2 | quantize: cap working memory size to avoid loading big tensors onto RAM (#27795) b10656 | ||
|
ShobhitandGitHub
|
cb300598d5 |
Feature: Added LIGHTNING_INDEXER support for Deepseek V4 ops on Vulkan Backend (#27453)
* vulkan: add LIGHTNING_INDEXER op
* vulkan: updated lightning_indexer.comp and ggml-vulkan.cpp with 128-lane dot-product reduction moved from a shared-memory tree to subgroupAdd.
* vulkan: cleanup; Skip bounds checks
* vulkan: cleanup FA_K_ONLY
* Revert "vulkan: cleanup FA_K_ONLY"
This reverts commit
b10655
|
||
|
Sigbjørn SkjæretandGitHub
|
1a946ec745 | pr2wt : use ssh/https remote in worktree depending on base (#27800) | ||
|
Xuan-Son NguyenandGitHub
|
fac889fb38 |
llama: model_loader: add TENSOR_READ_LAZY (#27794)
* llama: model_loader: add TENSOR_GET_ROW_LAZY * add --tensor-read-lazy * rename to TENSOR_READ_LAZY * gen docs * address commentsb10653 |
||
|
Aleksander GrygierandGitHub
|
cae63579b6 |
ui: Improve Chat Form Actions UI/UX (models selector, add panel) (#27746)
* ui : strip trailing container-format segments from parsed model names * ui : show reasoning and modality icons on model options and search by modality * ui : keep reasoning submenu visible regardless of model state * ui : add show-org-name-in-trigger display setting * ui : move model list into a submenu within the model selector * ui : make model option hover and focus highlight override the active state * ui : add raw model id tooltip to model selector options * feat: Enable microphone input as default for audio models * ui : fix eslint issues in chat form and model selector * ui: show modality icons instead of file submenu in chat add menu Assisted-by: pi * chore: Format * chore: Format * ui: add ModelCapability enum and shared modality/capability icon constants Assisted by: pi:GLM-5.3-Flash * ui: derive modality badge icons and labels from shared constants Assisted by: pi:GLM-5.3-Flash * ui: split model option icons into capabilities and modalities Replace the supportsThinking flag on ModelId with a capabilities object keyed like ModelModalities, so future capabilities (tool calls, etc.) slot in alongside reasoning. Icons and labels now come from the shared CAPABILITY_ICONS/MODALITY_ICONS constants. Assisted by: pi:GLM-5.3-Flash |
||
|
Kartik GuliaandGitHub
|
bcb6084a4e |
convert : fix Nemotron-H LoRA GGUF conversion (#27356)
* convert: fix Nemotron-H LoRA GGUF conversion * Removed redundant JSON import. |
||
|
Aleksander GrygierandGitHub
|
fe235f4343 |
ui: Replace per-conversation MCP overrides with per-conversation tool policy (#27745)
* ui: replace per-conversation MCP overrides with per-conversation tool policy MCP server enabled state is now global (server.enabled); per-conversation control moves to disabled tool keys and categories seeded into each new conversation. Aligns the add sheet with the dropdown options and flattens MCP tool groups in the tools submenu. Assisted-by: pi * ui: keep tool policy migration running when defaults parse fails A corrupt disabledToolKeys localStorage entry no longer aborts the migration; it falls through with empty defaults so legacy MCP server overrides still get converted. Assisted-by: pi * ui: fall back to global defaults when agentic flow has no tool policy Passing empty disabled sets bypassed the global defaults and could enable tools for callers that do not pass a policy yet. Assisted-by: pi * ui: align preferences section headers with their methods The Reasoning Effort and Working Directory headers sat above tool policy methods; move them above setCwd and setReasoningEffort. Also clarify the disabled tools JSDoc: existing rows with an unset field have an empty policy, defaults apply only when there is no active conversation. Assisted-by: pi * ui: gate MCP server avatars on conversation tool policy Servers whose tools are disabled for the current conversation (MCP category or server-scoped key) no longer show as enabled for the chat. Assisted-by: pi * ui: drop unused MCP category toggle from tools panel hook Per-conversation MCP control is server-granular; no component renders a whole-category toggle, so remove the dead API. Assisted-by: pi * ui: skip MCP init when flow policy disables the MCP category Resolve the effective tool policy before deciding whether to initialize MCP so flows that will not send any MCP tools skip the init work. Callers without a policy keep falling back to global defaults. Assisted-by: pi * chore: format * ui: restore reasoning section in mobile add sheet The sheet rewrite dropped it; the desktop dropdown still has it. MCP Prompts and Resources stay out of the sheet on purpose. Assisted-by: pi * ui: clear MCP server group key in enableAllToolsForServer The group key disables every tool of the server regardless of per-tool keys, so re-enabling a server from Settings did nothing while it was set. Assisted-by: pi * ui: skip MCP init when no policy-enabled server remains Extends the category-level check: the flow also skips MCP init when every globally-enabled server has its server-scoped group key disabled in the tool policy. Assisted-by: pi * ui: make Settings tools tab edit defaults with category toggles Adds per-category checkboxes and a caption stating the tab applies to new conversations; tool picks inside a chat only affect that chat. Assisted-by: pi * ui: gate cwd picker and mention picker on effective tool policy Both checked the global disabled set directly, so a conversation that disabled file_search still showed search as available. Assisted-by: pi * ui: clean up tool key helpers and store docs Documents getEnabledToolsForLLM properly, unstacks the JSDoc at isEntryEnabled, makes setToolEnabled persist like setCategoryEnabled (toggleTool now delegates to it), and routes the serverId-less MCP branch of toolKey through getMcpServerToolsKey so both key formats come from one place. Preferences banner comments become plain comments so they no longer read as class member docs. Assisted-by: pi * ui: indeterminate group checkboxes and inert grayed rows A category that is on with nothing enabled under it now shows the mixed checkbox state instead of a checked box next to 0/N. Rows grayed out by a disabled parent no longer stay clickable behind opacity. Assisted-by: pi * ui: gate MCP prompt and resource capabilities on tool policy hasPromptsCapability and hasResourcesCapability accept an optional set of usable server ids; ChatFormActions resolves it from global enablement minus the active conversation's policy. Restores the per-chat gating the old mcpServerOverrides provided; callers without arguments keep global behavior. Assisted-by: pi * ui: remove unmounted MCP submenu component Never rendered anywhere; its entries are duplicates (prompts and resources live in the attachment menu, servers in the add menu and sheet) that would need capability wiring maintained for nothing. Assisted-by: pi * ui: fix model information dialog width on all screen sizes The dialog sets container-type: inline-size, so auto width ignores its contents and collapses to padding. Give it an explicit viewport width on mobile and cap at 60rem on desktop. Assisted-by: pi * ui: scroll wide chat template in model information dialog Long unbreakable Jinja tokens blew out the table and dialog width; the block now scrolls horizontally instead of stretching. Assisted-by: pi * ui: use fixed table layout in model information dialog Auto table layout sizes columns to content min-content, so the chat template's long lines kept inflating the dialog despite the scroll wrapper. Fixed layout pins the first column and gives the value column a definite width the wrapper can scroll within. min-w-0 on the grid item guards the same path on the grid side. Assisted-by: pi * ui: make model information dialog full-screen on mobile Matches the settings dialog pattern: full viewport below md, calc-sized and capped at 60rem on desktop. Assisted-by: pi * ui: stack chat template row in model information dialog Label above the block in a single full-width cell, so the template gets the whole table width and its horizontal scroll is usable on narrow screens. Assisted-by: pi * ui: scroll model information header with the content The base dialog header is sticky; this dialog overrides it to relative so the title and description scroll away with the body. relative keeps the header as the close button's containing block. Assisted-by: pi * ui: replace literal comment text in sheet group snippet A // line inside the Svelte snippet rendered as visible text; use an HTML comment. Assisted-by: pi * ui: let indeterminate state win over checked in group checkboxes The checkbox indicator snippet renders the check icon whenever checked, so the mixed state never showed. Pass the checked prop as false while indeterminate. Assisted-by: pi * ui: initialize only policy-enabled MCP servers for a flow ensureInitialized accepts an optional server id set; the agentic flow passes the servers its tool policy leaves usable, so servers disabled for the conversation no longer get connected. Callers without arguments keep the global behavior. Assisted-by: pi * ui: derive group checkbox state in useToolsPanel Moves the mixed-state derivation out of the submenu and sheet snippets into one getGroupCheckState accessor; the snippets just consume checked and indeterminate. Assisted-by: pi * ui: gate /prompt command on the conversation tool policy The slash command's availability now follows the same rule as the agentic flow instead of the global capability check, so it disables itself when the conversation's policy leaves no usable MCP server. Assisted-by: pi * ui: remove dead MCP prompt menu trigger chain The /prompt slash command is the surviving trigger; the menu-button path (onMcpPromptClick, hasMcpPromptsSupport, showMcpPromptButton, the MCP_PROMPT attachment item and its unrendered item arrays) has no consumer left. Message display for inserted prompts is untouched. Assisted-by: pi * ui: render dash for mixed-state group checkboxes The accessor refactor dropped the checked-and-not-indeterminate guard, so the category-on flag won and the dash never showed. The tooltip keeps using the raw parent flag since clicking a mixed group still disables it. Assisted-by: pi * ui: fix group checkbox sticking checked after disable Clicking a mixed-state group box let bits-ui optimistically flip its internal checked flag; the derived checked prop did not change across the transition (both mixed and off map to checked=false), so Svelte never applied the settled value and the check icon stuck while the count already read 0/7. Pass the parent flag as checked and the mix as indeterminate, so every group toggle changes checked; render the dash on top of a checked box for the mixed state. Assisted-by: pi * fix: UI for Model Information dialog * ui: keep MCP connections stable across policy switches ensureInitialized folds the policy into its config signature, so alternating two conversations with different policies tore down and reconnected every server with health checks included. Tool collection already filters by the flow policy, so initialize every settings-enabled server instead and never pass a policy into the MCP config. The duplicated policy-server check becomes one accessor on ConversationPreferences. Assisted-by: pi * ui: remove dead MCP resources menu trigger chain Same shape as the earlier prompt trigger cleanup: nothing renders the MCP resources menu button, and the only live entry into resource browsing is Settings > MCP Servers plus the attachment resource picker. Drop onMcpResourcesClick, hasMcpResourcesSupport, MCP_RESOURCES_CLICK, the AttachmentItemVisibleWhen enum and hasResourcesCapability; the resources display, browser and picker components are untouched. Assisted-by: pi |
||
|
Gaurav GargandGitHub
|
2bb9bddafa |
spec: Add benchmark-only synthetic speculative acceptance options (#27711)
* Add benchmark-only synthetic speculative acceptance to llama-server and llama-cli * Address review comments * Address review comments * Add some comments in the codeb10649 |
||
|
|
deae5ee133 |
model : simplify MiniMax-01 graph (#27790)
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>b10648 |
||
|
Xuan-Son NguyenandGitHub
|
f29551215b |
args: add --video-* CLI arguments (#24318)
* args: add --video-* CLI arguments * gen docs * nits * add mtmd_helper_init_optb10647 |
||
|
Niklas WenzelandGitHub
|
915dc6d38c | metal : fix memory leaks due to missing autoreleasepools (#27758) b10646 | ||
|
Jonas JandGitHub
|
c5fc7e3488 |
llama : add --n-cpu-ffn option (#26622)
* common : dedupe --n-cpu-moe / --spec-draft-n-cpu-moe override loops * common : add --n-cpu-ffn to CPU-offload dense FFN weights of first N layers * common : generalize llm_ffn_block_regex over the FFN regex, drop TODOb10645 |
||
|
|
d7a2074112 |
models : support nanbeige4.2-3B (#27730)
Co-authored-by: admin <lizongqiang@kanzhun.com>b10644 |
||
|
Max KrasnyanskyandGitHub
|
192067b72d |
hexagon: support for multi-NPU devices (IQ9, IQ10) and fully asynchronous backend (#26501)
* hexagon: use non-host bufs by default and make the backend fully async * hex-hb: remove optional hostbuf support and fix async copy * hex-unary: relax supported unary check * hex-bufs: use same get_alignment for host bufs * snapdragon: bump android_platform to 34 * hex-rows: super hacky get/set rows for q8_0 * hex-get-rows: fix q8_0 * hex-get-rows: supprot for f16 and cleanup for q8_0 * hex-get-rows: generic macros and specialized thread funcs * hex-get-rows: add DMA pipeline, vtcm_layout and kernel params * hex-set-rows: fix q8_0 support, add dma and tracing * hex-tests: override nmse threshold for HTP of Q8_0 quants * hex-fa: add support for Q8_0 with inplace dequantizers * hex-get-rows: simplify type dispatch * hex-rows: simplify GET/SET_ROWS DMA pipeline * hex-async: add events, set/get-tensor-async and rest of the async api support * hex-repack: use slice instead of expert in repack functions * hex-cpy: update event/async-cpy logging * hex-set-rows: optimize smaller tensors * hex-geglu: fix perf regression with larger tensors * hex-get-rows: add missing header * hex-set-rows: add missing header * hex-bufs: ressurect GGML_HEXAGON_HOSTBUF but disable it by default * hexagon: do not reject ops with non-heaxon buffers * hex-get-rows: apply >=32 restriction only for q8_0 * hex-res: bump vtcm acquire timeout to 10 seconds * hex-bufs: add support for cloning buffers between sessions to speed up tensor copies * hex-async: rework event recording and batch flushing and integrate with meta backend * hex-bufs: improved handling of repacked tensors * hex-repack: handle get_tensor_2d offsets * hex-dev: add support for devices with multiple NPUs * hex-sync: add support for sync tokens to synchronize npu devices for async splits * hex-mmap: cleanup mmap calls and add a retry for robustness * hex-sync: add failsafe if sync wait gets stuck * hex-sync: use sync_seq to check for completed events * hex-sync: rotate tokens for extra robustness * hex-devs: add supprot for legacy device names for now * hex-bufs: add support for auto-cloning buffers from diff sessions * hex-fusion: simplify and optimize htp-opnode fusion handling * hex-sync: override opnode name so that it shows up in the profiles * hex-trace: update scripts to handle multiple devices * hex-sync: bump the size of the opbatch queue and number of sync tokens * hex-cpy-sync: do not explicitly flush opbatches in cpy_tensor_async and add support for cpy-dma * hex-sync: add graph-flush threshold to avoid single op batches * hex-sync: add sync_peer so that we can flush peers we depend on during cross-device ops * hex-bufs: introduce tensor->extra and shadow_bufs for repacking * hex-l2: flush tiny tensors inline * hex-sync: use explicit l2flush for sync tokens * hex-extra: track weight flags via tensor extra * hex-fence: rename sync to fence * hex-repack: proper handling of set-tensor-2d in the shadow_buf * hex-trace: remove obsolete opstage mask that we used for profiling * hex-env: remove obsolete use_hmx variable * hexagon: new unified run.py and build.py and updated docs * snapdragon: update run script to auto-escapt test-backend-op -p argument * hex-scripts: fix trailing spaces * hex-scripts: fix flake8 warnings * snapdragon: cleanup dst lib/bin dirs before copying new build * hex-ops: add support for allreduce * hex-ar: improved allreduce with dma pipeline * hex-ar: align macros * hex-ar: consistent use of fence_seq * hex-ar: add AR_SELECT env var to select ALLREDUCE kernel or fallback * hex-ar: add proper synchronize handling for ALLREDUCE * hex-opbatch: looks like we now just rely on backend.synchronise to flush the batches, no need to flush them by threshold * hex-ar: bump block size to improve dma efficiency * hex-ar: fused ALLREDUCE+ADD * hex-ar: cleaner fence buffer management * hex-ar: futher allreduce tweaking to remove race conditions * hex-ar: add simple solver and remove non-dma kernels * hex-ar: add row-broadcast to fuse with bias ADD * hex-fence: pass seq numbers via op_params * hex-ar: allow for both entry/exit seq for completing entry wait * hex-ar: align macros * hex-ar: do not refetch broadcast row * hex-fusion: move all fusion into opbatch::add_op for consistency with ALLREDUCE and things * hex-fusion: fix incorrect MUL_MAT reordering * hex-mm: make fused 2x and 3x matmuls more generic * hex-fusion: move tensor fusion tagging to graph_compute * hexagon: make sure to copy tensor->extra by value * hex-get-rows: fix offset calc with row-chunking * hex-repack: get_tensor_2d fixes for non-zero offsets * snapdragon: make profile/trace scripts more robust and donot mix stdout/stderr by default * hex-devices: use legacy device nameing by default to ease the transition * hex-devices: hardcode CDSP domain IDs for current devices for now * hex-optrace: improve multi-NPU timestamp alignment and overall handling of cycle values * hex-optrace: more robust handling of the fence eventsb10643 |
||
|
Xuan-Son NguyenandGitHub
|
925e117994 |
llama: add token ID tracking to KV cell (#27762)
* kv: track token id * rm get_prev_tokens, move it to the main pr * nits * add get_prev_tokensb10642 |
||
|
Aleksander GrygierandGitHub
|
539f24529b |
ui: Move Settings and MCP Servers routes to dialog-based views (#27744)
* ui : open MCP servers in a dialog from the chat form Replace the MCP servers submenu with a single "MCP Servers" item that opens a new DialogMcpServers dialog instead of navigating to the /mcp-servers route. Assisted-by: pi * ui : browse MCP resources from the server card Make the Resources capability badge clickable so it opens the MCP resources browser dialog, and drop the page-only chrome from SettingsMcpServers. Assisted-by: pi * ui : remove mcp-servers route and sidebar entry MCP servers are now managed in a dialog, so drop the dedicated route and the sidebar icon that navigated to it. Assisted-by: pi * ui : remove unused MCP servers submenu component The submenu was replaced by the MCP servers dialog, so delete the component and its export. Assisted-by: pi * feat(ui): add DialogSettingsChat dialog * refactor(ui): switch SettingsChat to in-app section navigation * feat(ui): open settings as dialog from sidebar * refactor(ui): remove settings route and URL-based settings navigation * fix(ui): adjust MCP dialogs for new base sizing * chore: Formatting & linting |