[Chore] Use upstream Transformers DeepseekV4Config by hmellor · Pull Request #61392 · vllm-project/vllm

hmellor · 2026-10-13T10:07:49Z

Overview

Split from #60823: deletes the vendored DeepseekV4Config and ports the DeepSeek-V4 code to the upstream Transformers 5.16.1 schema. Also updates the transformers-upgrade skill.

Related PRs:

[Chore] Port more vendored configs and processors to upstream Transformers #60823: the larger unvendoring PR this is split from; DeepSeek-V4 has been dropped from it.
[Bugfix][DeepseekV4] Harden compress_ratio fallback for transformers >=4.57 #43443: patches one reader of the popped compress_ratios (attention.py). This PR covers every reader, so it supersedes [Bugfix][DeepseekV4] Harden compress_ratio fallback for transformers >=4.57 #43443.

Claims

Deletes vllm/transformers_utils/configs/deepseek_v4.py. Upstream folds compress_ratios and num_hash_layers into layer_types / compress_rates / mlp_layer_types:
- V4-only code (attention, decoder, CPU/XPU/NVIDIA MoE, the C4/C128 CuTe DSL kernels) reads the upstream fields directly.
- Code that also sees DeepSeek-V4.1 configs (the shared MLA backends, mla_attention, the arch convertor, the spec-decode proposer and the warmup kernels) uses get_compress_ratios. That helper returns V4.1's legacy field, or derives it from the upstream fields with the MTP layers appended. V4.1 can't move to the upstream schema: it uses compress ratios 1 and 2, which have no upstream layer type.
mm_prefix_span_leading_pad_modulus moves from the config onto DeepseekV4ForConditionalGeneration, where the model runner already looks for it. The sparse-SWA image-span widening is now keyed on an upstream DeepseekV4Config with vision_n_layers > 0, matching what the vendored config set.
is_rope_parameters_nested accepts V4's {"main", "compress"} rope layout.
No numerics change for DeepSeek-V4-Flash: greedy outputs are bit-identical to main, and GSM8k, GPQA Diamond and MTP acceptance match.
transformers-upgrade skill:
- Models in vllm/models/ must run GSM8k and GPQA Diamond against main, and these evals should disable FlashInfer autotuning.
- Adds two hazards: undeclared fields with no default, and top-level configs that no longer forward reads to text_config.
- Adds a re-grep for old field names after rebasing.
- Drops the PR and issue numbers from the skill.

Validation

deepseek-ai/DeepSeek-V4-Flash on 8x B200, main (bc3ddb198a) on GPUs 0-3 vs this branch on GPUs 4-7, TP4 each, transformers 5.19.0.

Check	`main`	This PR
GSM8k, 5-shot, 1319 questions, autotune off	0.944, 0.944	0.937, 0.947
GSM8k with MTP (1 token), autotune off	0.947	0.946
MTP draft acceptance	95.5% (57174 / 59896)	95.4% (57041 / 59771)
GSM8k with MTP, rerun on `84f38ff`	0.948	0.949
MTP draft acceptance, rerun on `84f38ff`	95.4% (57236 / 59977)	95.5% (57494 / 60209)
GPQA Diamond, 8 repeats, reasoning effort high	0.538	0.534
Greedy decode, 4 short prompts x 64 tokens, autotune off		bit-identical logprobs

FlashInfer re-tunes its FP4 MoE tactics on every server start, even when given a cache. With autotuning on, the two servers picked different tactics and GSM8k drifted by about 1 point (0.947-0.951 vs 0.939-0.941). With --kernel-config '{"enable_flashinfer_autotune": false}' greedy outputs match exactly, so the drift is tactic selection, not this change.

Config dumps for Flash, Flash-Vision-Exp and Pro, using the skill's scripts: get_compress_ratios, the per-layer upstream reads, the hash-layer count and the C4/C128 kernel flags all reproduce each checkpoint's legacy values exactly; the mm-prefix clamp is on only for Vision-Exp, as before; and the DeepSeek-V4 RoPE cos_sin_cache is bit-identical for sliding, CSA and HCA layers. Vision-Exp and Pro were not served (not enough disk on the host); Vision-Exp's config.json sets every vision_* field the processor reads.

vllm serve deepseek-ai/DeepSeek-V4-Flash --tensor-parallel-size 4 --kv-cache-dtype fp8 \
  --max-model-len 65536 --tokenizer-mode deepseek_v4 --reasoning-parser deepseek_v4 \
  --kernel-config '{"enable_flashinfer_autotune": false}' \
  [--speculative-config '{"method": "mtp", "num_speculative_tokens": 1}']
python tests/evals/gsm8k/gsm8k_eval.py --port <port> --num-shots 5 --max-tokens 512
OPENAI_API_KEY=EMPTY python tests/evals/gpt_oss/run_gpqa_eval.py --model deepseek-ai/DeepSeek-V4-Flash \
  --reasoning-effort high --base-url http://localhost:<port>/v1 --n-threads 128 --max-output-tokens 32768

pytest tests/transformers_utils tests/config tests/tokenizers_ \
  tests/models/multimodal/processing/test_deepseek_v4_vl.py \
  "tests/test_config.py::test_draft_inherits_ep_only_for_moe"
# main: 1 failed, 1137 passed; this PR: 1 failed, 1139 passed
# the failure in both is test_mp_reducer, which starts an engine (GPUs hidden for this run)

Pull Request Checklist

I used vLLM's /pr-checklist skill. (Mandatory for agents, optional for humans).
AI assistance was used during the creation of this PR.
Design Fit: Minimizes impact on core components, reuses existing functionality, and justifies added complexity.
Testing and Validation: Validates the change and ensures any added tests are meaningful and reliable, with CI coverage or documented CI resource constraints and validation performed outside CI.
Code Quality and Style: Keeps code and comments clear and concise, and updates relevant documentation and examples.
Pull Request Contents: Includes a brief summary and relevant links, supports claims with evidence, explains root causes and implementation trade-offs, and follows the contributing guide.

🤖 Generated with Claude Code

Split from vllm-project#60823. Derives the legacy compress_ratios, num_hash_layers and mm-prefix flags from the upstream layer_types / mlp_layer_types fields. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

…ng, evals for vllm/models Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

Keep get_compress_ratios only where V4.1 configs also flow, inline the mm-prefix clamp check and drop PR numbers from the skill. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

claude

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

hmellor · 2026-10-14T13:05:58Z

/ci run --allow-stale

github-actions · 2026-10-14T13:06:21Z

✅ Triggered Buildkite CI #94801 for commit 84f38ff455a6.

⚠️ This PR is 101 commits behind upstream main. Running CI at your own risk because --allow-stale was requested; outdated CI configuration may cause failures. Before merging, merge or rebase onto the latest main, then rerun /ci run on the latest PR commit.

hmellor · 2026-10-14T13:07:55Z

/ci run

github-actions · 2026-10-14T13:08:18Z

✅ Triggered Buildkite CI #94802 for commit 97cef856b4a1.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

…dor-deepseek-v4

hmellor · 2026-10-14T16:17:33Z

/ci run

github-actions · 2026-10-14T16:18:05Z

✅ Triggered Buildkite CI #94855 for commit 88bdb7a60a7e.

ℹ️ This PR is 11 commits behind upstream main. Merged upstream main into this branch as 88bdb7a60a7e before starting CI. Pull the branch before pushing more changes.

hmellor · 2026-10-14T22:39:57Z

/ci retry

github-actions · 2026-10-14T22:40:15Z

✅ Queued 1 failed job(s) for retry in Buildkite CI #94855.

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Kevin H. Luu <khluu000@gmail.com>

### What this PR does / why we need it? Port `vllm-ascend` to vLLM main `c0b6f74` (bumped from `a98247a`) while keeping the `v0.31.0` release lane green. The upstream window moved the KV-cache planning helpers into a new module, made EAGLE draft groups explicit, typed DeepSeek-V4 configs through upstream Transformers, removed Medusa/MLP speculator support, and changed several spec-decode/KV-cache/EPLB signatures. Every gate below is either a `vllm_version_is("0.31.0")` split or an attribute/argument probe where two main dev builds share one version string. | Files | Upstream vLLM change (why it breaks vllm-ascend) | vllm-ascend adaptation | |-------|--------------------------------------------------|------------------------| | `.github/vllm-main-verified.commit` | — (main-lane pin only) | Verified main commit bumped `a98247a` -> `c0b6f74` | | `tests/e2e/pull_request/four_card/test_minimax_m3.py` | The main-lane pin re-JITs the static kernels, so the first generate on a cold cache shifts the near-uniform dummy logprobs (~3e-3) enough to flip a greedy token | On main run a throwaway `generate_greedy_logprobs` before the baseline/replay pair so both measure the warmed steady state and share the cached prefix; v0.31.0 keeps the single-shot comparison | | `tests/ut/core/test_circular_buffer.py` | [vllm#61126](vllm-project/vllm#61126) (cascade-attention removal) moved the prefix-block query up to `KVCacheManager` and dropped it from single-type managers | On main assert only the manager's own `get_num_skipped_tokens(65536)`; keep the combined `get_num_common_prefix_blocks(...)` assert on v0.31.0 | | `tests/ut/core/test_kv_cache_interface.py` | [vllm#61183](vllm-project/vllm#61183) moved `is_kv_cache_spec_uniform` into `vllm.v1.core.kv_cache_planning` | Version-aware import (`kv_cache_utils` on v0.31.0 / `kv_cache_planning` on main) | | `tests/ut/distributed/ascend_store/_mock_deps.py` | [vllm#61183](vllm-project/vllm#61183) added the `vllm.v1.core.kv_cache_planning` module | Add it to the mocked-vLLM module list | | `tests/ut/distributed/ascend_store/test_request_lifecycle.py` | [vllm#58668](vllm-project/vllm#58668) trims `num_computed_tokens` to the tokens actually kept when a request stops mid token-list | On main subtract `(accepted + 1) - <newly appended output tokens>` from the expected computed count; keep the old expectation on v0.31.0 | | `tests/ut/distributed/kv_transfer/kv_p2p/test_mooncake_connector.py` | [vllm#61183](vllm-project/vllm#61183) moved `get_kv_cache_config_from_groups` / `is_kv_cache_spec_uniform` | Version-aware import | | `tests/ut/models/test_deepseek_v41_cache.py` | [vllm#61183](vllm-project/vllm#61183) relocated the planning helpers; [vllm#53558](vllm-project/vllm#53558) gave `_max_memory_usage_bytes_from_groups` a leading `builder` argument | Import from the live planning module and probe the upstream signature to call the 2-arg (v0.31.0) vs 3-arg (main) form, feeding an Ascend `_get_pool_bytes_per_block` builder | | `tests/ut/ops/test_gdn_attn_builder.py` | [vllm#58737](vllm-project/vllm#58737) derives `num_reqs` from the block table and asserts its row count equals `seq_lens`; Ascend's `unpadded()` keeps the graph-padded rows | New UT: a padded 4-row block table with 2 real requests must still yield a 2-row `cache_indices` | | `tests/ut/patch/platform/test_packed_kv_cache_compat.py` | [vllm#61183](vllm-project/vllm#61183) relocated the patched module; [vllm#61455](vllm-project/vllm#61455) removed `_annotate_eagle_groups` from main | Patch `cache_patch._kv_cache_planning`; new UT asserts the packed hook tolerates an upstream module without the EAGLE-annotation API | | `tests/ut/patch/platform/test_patch_eplb.py` | [vllm#57549](vllm-project/vllm#57549) emits the async EPLB cycle-committed log from the async worker wrapper on main | Expect the workspace-move log only on v0.31.0 | | `tests/ut/patch/platform/test_patch_kv_cache_coordinator.py` | [vllm#61183](vllm-project/vllm#61183) relocated the patch entry point | Assert against `kcu_patch._kv_cache_planning` instead of `vllm.v1.core.kv_cache_utils` | | `tests/ut/patch/platform/test_patch_pp_mtp.py` | [vllm#56322](vllm-project/vllm#56322) makes `ModelConfig.verify_with_parallel_config` read `parallel_config.enable_hc_sp` | Add `enable_hc_sp=False` to the constructed parallel config (ignored on v0.31.0) | | `tests/ut/patch/platform/test_prefix_cache_cp_patches.py` | [vllm#60043](vllm-project/vllm#60043) removed the `mamba_cache_mode="all"` leftover and requires `"align"` with prefix caching; [vllm#53558](vllm-project/vllm#53558) added the `builder` argument | Version-select the mamba cache mode; probe the upstream admission signature for the `builder` argument | | `tests/ut/spec_decode/test_mla_dspark.py` | [vllm#60692](vllm-project/vllm#60692) renamed the capture hook `precompute_context_kv(num_reqs)` -> `prepare_forward(num_reqs, attn_metadata)` | Version-split the expected capture args/kwargs | | `tests/ut/test_compressed_prefix_cache.py` | [vllm#61183](vllm-project/vllm#61183) moved `is_kv_cache_spec_uniform` | Version-aware import | | `tests/ut/worker/test_attn_utils_v2.py` | [vllm#61455](vllm-project/vllm#61455) annotates EAGLE groups from `spec.is_draft`; [vllm#60766](vllm-project/vllm#60766) resolves pin memory lazily via `is_pin_memory_available()` | New UT that draft-layer specs get tagged; on main monkeypatch `is_pin_memory_available` to `False` | | `tests/ut/worker/test_model_runner_v2.py` | [vllm#56716](vllm-project/vllm#56716) suppresses EPLB even for `skip_eplb` dummy runs; [vllm#60946](vllm-project/vllm#60946) renamed the PCP staging hook | Add a `runner.eplb` stub; call the version-appropriate staging method | | `tests/ut/worker/test_mtp_pcp_speculator_v2.py` | [vllm#60885](vllm-project/vllm#60885) rejects a mismatched explicit float16/bfloat16 KV-cache dtype | Build the model config with `bfloat16` instead of `float32` | | `tests/ut/worker/test_pcp_manager_v2.py` | [vllm#60946](vllm-project/vllm#60946) renamed `prepare_inputs_to_capture` -> `prepare_inputs_for_capture_or_profile` | Patch and call the live method name per lane | | `tests/ut/worker/test_v2_dummy_padding.py` | [vllm#61767](vllm-project/vllm#61767) collects DP votes through `get_dp_group().cpu_all_reduce_()` instead of `dist.all_reduce(..., group=cpu_group)` | Version-split the mocked collective | | `tests/ut/worker/v2/test_eplb_controller.py` | [vllm#57549](vllm-project/vllm#57549) gives `add_model` an optional `model_name` and names the draft entry | Expect the 3-arg call (`model_name="draft (draft)"`) on main, 2-arg on v0.31.0 | | `vllm_ascend/_310p/worker_310p.py` | [vllm#61183](vllm-project/vllm#61183) moved `get_kv_cache_groups` | Version-aware import | | `vllm_ascend/distributed/ec_transfer/ec_memcache_connector.py` | [vllm#57904](vllm-project/vllm#57904) dropped `_get_encoder_cache_hidden_dim` from `cpu.common` (encoder widths are measured at startup) | On main size the receive buffer as `out_hidden_size * (1 + len(deepstack_visual_indexes))` (Qwen3-VL DeepStack) else `get_inputs_embeds_size()`; keep the helper on v0.31.0 | | `vllm_ascend/models/deepseek_v4/compressor.py` | [vllm#61392](vllm-project/vllm#61392) switched `DeepseekV4Config` to the upstream Transformers class | Import from `transformers` on main, from `vllm.transformers_utils.configs.deepseek_v4` on v0.31.0 | | `vllm_ascend/models/deepseek_v4/indexer.py` | Same [vllm#61392](vllm-project/vllm#61392) | Same version-aware import | | `vllm_ascend/models/deepseek_v4/mm_preprocess.py` | Same [vllm#61392](vllm-project/vllm#61392) | Same version-aware import | | `vllm_ascend/models/deepseek_v4/model.py` | [vllm#61392](vllm-project/vllm#61392) moves the config to Transformers, folds `compress_ratios`/`num_hash_layers` into `layer_types`/`mlp_layer_types`, and keys YaRN under the compressor (`compress`) group | Restore the legacy fields on main from the new lists, read the `compress`-scoped rope params with safe defaults, and keep the legacy path on v0.31.0 | | `vllm_ascend/models/deepseek_v41/cache_config.py` | [vllm#61183](vllm-project/vllm#61183) moved `may_override_num_blocks` | Version-aware import | | `vllm_ascend/models/glm5next/cache_config.py` | [vllm#61183](vllm-project/vllm#61183) moved `create_kv_cache_group_specs` / `may_override_num_blocks` | Version-aware import | | `vllm_ascend/ops/gdn_attn_builder.py` | [vllm#58737](vllm-project/vllm#58737) derives `num_reqs` from the block table and asserts row parity with `seq_lens` | On main trim the padded block table to `seq_lens.shape[0]` before `mamba_get_block_table_tensor` when the sizes differ | | `vllm_ascend/patch/__init__.py` | [vllm#61183](vllm-project/vllm#61183) moved `get_kv_cache_config_from_groups` to `kv_cache_planning` | Update the patch-registry doc | | `vllm_ascend/patch/platform/__init__.py` | [vllm#61183](vllm-project/vllm#61183) moved the KV-cache planning/patcher module | Update the patch-registry doc | | `vllm_ascend/patch/platform/patch_eplb.py` | [vllm#57549](vllm-project/vllm#57549) moves the async EPLB cycle-committed log onto the async worker path on main | Report completion from the wrapped async worker (rank 0) on main; keep the workspace-move log on v0.31.0 | | `vllm_ascend/patch/platform/patch_kv_cache_utils.py` | [vllm#61183](vllm-project/vllm#61183) moved the planning helpers; [vllm#53558](vllm-project/vllm#53558) added the `builder` argument; [vllm#61455](vllm-project/vllm#61455) dropped `_annotate_eagle_groups` | Resolve the live planning module by version, patch its symbols, gate the EAGLE annotation on attribute presence, and dispatch the admission helper by argument count | | `vllm_ascend/quantization/methods/kv_cache/turboquant/cache.py` | [vllm#61183](vllm-project/vllm#61183) moved `may_override_num_blocks` | Version-aware import | | `vllm_ascend/spec_decode/__init__.py` | [vllm#61701](vllm-project/vllm#61701) removed Medusa/MLP speculator support | Import `AscendMedusaProposer` only on v0.31.0; raise a clear error for `method="medusa"` on main | | `vllm_ascend/spec_decode/medusa_proposer.py` | [vllm#61701](vllm-project/vllm#61701) removed `vllm.v1.spec_decode.medusa` | Add `# type: ignore[import-not-found]`; the module only loads on the lane that still ships it | | `vllm_ascend/worker/v1/model_runner.py` | Same [vllm#61701](vllm-project/vllm#61701) | Gate the Medusa import and the `isinstance(self.drafter, AscendMedusaProposer)` branch on v0.31.0 | | `vllm_ascend/worker/v2/attn_utils.py` | [vllm#61455](vllm-project/vllm#61455) annotates EAGLE groups from `spec.is_draft` | On main tag the spec with `is_draft` from `attn_module.is_draft_layer` | | `vllm_ascend/worker/v2/pcp_manager.py` | [vllm#60946](vllm-project/vllm#60946) added the `moe_gathers_tokens` argument and renamed `prepare_inputs_to_capture` -> `prepare_inputs_for_capture_or_profile` | Forward the new argument; version-split the staging hook | | `vllm_ascend/worker/v2/sample/apply_top_k_top_p.py` | [vllm#61996](vllm-project/vllm#61996) added `has_topp_only_rows` to the split top-p pipeline | Accept and ignore the extra argument on main (the PyTorch fallback handles top-p-only rows) | | `vllm_ascend/worker/v2/spec_decode/dflash/aclgraph.py` | [vllm#60692](vllm-project/vllm#60692) renamed `precompute_context_kv` -> `prepare_forward` | Pass the capture hook the installed vLLM expects | | `vllm_ascend/worker/v2/spec_decode/lmhead_tp_utils.py` | [vllm#59554](vllm-project/vllm#59554) added `num_steps` to `Speculator.sample_draft` | Version-split the override signature and forward `num_steps` on main | | `vllm_ascend/worker/v2/spec_decode/rejection_sampler_utils.py` | [vllm#61451](vllm-project/vllm#61451) renamed the logits-stats kernel argument `num_speculative_steps` -> `num_logits` and moved bonus-token detection into `_is_bonus_logit` | Pass `num_logits` on main; keep `num_speculative_steps` on v0.31.0 | | `vllm_ascend/worker/worker.py` | [vllm#61183](vllm-project/vllm#61183) moved `get_kv_cache_groups`/`unify_hybrid_kv_cache_specs`; [vllm#57516](vllm-project/vllm#57516) reclaims cached memory via `torch.accelerator.empty_cache()` inside `load_model` | Version-aware imports; on main neutralize `torch.npu.empty_cache` while loading weights under the sleep-mode pluggable memory pool | ### Does this PR introduce _any_ user-facing change? No. This is an internal main2main compatibility update; on the `v0.31.0` release lane behavior is unchanged, and on main the only behavioral note is that Medusa speculative decoding now raises a clear error (upstream removed it in vllm#61701). ### How was this patch tested? - Unit tests: the changed UT files above (KV-cache planning/coordinator, EPLB, PCP/spec-decode, DeepSeek-V4/V4.1 cache, mamba align, `worker/worker.py`). - E2E: `tests/e2e/pull_request/four_card/test_minimax_m3.py`. - Run the main lane against the pinned `c0b6f74` and the `v0.31.0` lane via `vllm_version_is`. --- ### Checklist #### Testing - [ ] **Boundary / edge-case coverage**: I reviewed whether this change involves boundary conditions (e.g. empty input, single/zero element, min/max values, off-by-one, sequence-length limits, integer overflow). Where applicable, I added unit tests covering these edge cases in `tests/ut/` or `tests/e2e/`. - [ ] **Regression test**: If this PR fixes a bug, I added a regression test that would fail without the fix. - [ ] **Error / failure path coverage**: I covered abnormal paths (exceptions, timeouts, out-of-range, division by zero) in addition to the happy path. - [ ] **Meaningful & reliable tests**: New tests assert real behavior and are not written merely to satisfy CI. - [ ] **Performance impact**: If this PR touches a performance-sensitive path, I attached benchmark results and added performance regression tests. - vLLM main: vllm-project/vllm@a98247a Signed-off-by: zhangxinyuehfad <starmoon_zhang@163.com>

hmellor and others added 3 commits October 13, 2026 08:45

[Skill] Transformers upgrade: undeclared fields, text_config forwardi…

4121e96

…ng, evals for vllm/models Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

[Skill] Disable FlashInfer autotune for main-vs-branch evals

250eb2b

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

mergify Bot added deepseek Related to DeepSeek models multi-modality Related to multi-modality (#4194) DSv4 DSv4.1 Related to DeepSeek-V4.1 models cpu Related to CPU backends speculative-decoding agents Related to information for agents labels Oct 13, 2026

cjackal marked this pull request as ready for review October 14, 2026 04:28

cjackal requested review from AndreasKaratzas, DarkLight1337, LucasWilkinson, MatthewBonanni, benchislett, dllehr-amd, hongxiayang, njhill, pavanimajety, tjtanaa, ywang96 and zyongye as code owners October 14, 2026 04:28

claude Bot reviewed Oct 14, 2026

View reviewed changes

hmellor mentioned this pull request Oct 14, 2026

[Chore] Port more vendored configs and processors to upstream Transformers #60823

Merged

5 of 6 tasks

Isotr0py approved these changes Oct 14, 2026

View reviewed changes

hmellor enabled auto-merge (squash) October 14, 2026 13:05

github-actions Bot added the ready ONLY add when PR is ready to merge/full CI is needed label Oct 14, 2026

Merge branch 'main' into unvendor-deepseek-v4

97cef85

hmellor and others added 2 commits October 14, 2026 17:15

Update DeepSeek-V4 test configs to upstream DeepseekV4Config fields

87b2832

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

Merge remote-tracking branch 'origin/unvendor-deepseek-v4' into unven…

6a94cd5

…dor-deepseek-v4

Merge branch 'main' into unvendor-deepseek-v4

88bdb7a

vllm-bot merged commit b965ee5 into vllm-project:main Oct 15, 2026
271 of 273 checks passed

hmellor deleted the unvendor-deepseek-v4 branch October 15, 2026 08:11

This was referenced Oct 15, 2026

[Chore] Use upstream Transformers MiniMaxM3VLConfig #61813

Open

[Chore] Use upstream Transformers InklingProcessor #61883

Merged

zhangxinyuehfad mentioned this pull request Oct 21, 2026

[Misc] vllm main2main 1017 vllm-project/vllm-ascend#19302

Merged

5 tasks

zkawfanx mentioned this pull request Oct 22, 2026

[Start here] Classify the pinned DSv4.1 timeline and prepare chronological backport batches zkawfanx/vllm#18

Open

3 of 5 tasks

Navigation Menu

Repository navigation

[Chore] Use upstream Transformers DeepseekV4Config - #61392

[Chore] Use upstream Transformers DeepseekV4Config#61392
vllm-bot merged 8 commits into
vllm-project:mainfrom
hmellor:unvendor-deepseek-v4

hmellor commented Oct 13, 2026 •

edited

Loading

Uh oh!

claude Bot left a comment

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

hmellor commented Oct 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Overview

Claims

Validation

Uh oh!

claude Bot left a comment

Choose a reason for hiding this comment

Claude Code Review

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

hmellor commented Oct 14, 2026

Uh oh!

github-actions Bot commented Oct 14, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

hmellor commented Oct 13, 2026 •

edited

Loading