[CRITICAL/base] Register Qwen3_5MoeForCausalLM in model registry + copy adapter to models/

WITHOUT THIS CHANGE: vllm cannot load Qwen3.6-35B-A3B model.
The model's config.json has architectures=['Qwen3_5MoeForCausalLM'],
but registry.py only had Qwen3ForCausalLM and Qwen3MoeForCausalLM.
Model init fails → ALL 50+ functional tests fail → zero competition score.

Changes:
1. registry.py: Add Qwen3_5MoeForCausalLM -> ('qwen3_5', 'Qwen3_5MoeForCausalLM')
2. Copy vllm_adapter/qwen3_5.py -> vllm/model_executor/models/qwen3_5.py
   so the registry's module resolution finds it.

The adapter (588 lines) implements:
- Qwen3_5MoeMLP, Qwen3_5MoeSparseMoeBlock (256 experts, top-8)
- Qwen3_5MoeAttention (with shared_expert support)
- Qwen3_5MoeDecoderLayer, Qwen3_5MoeModel, Qwen3_5MoeForCausalLM
- All imports use absolute paths (from vllm.xxx) + relative (.interfaces)
  which work correctly from vllm/model_executor/models/ directory.

CCCL context: agent_rle.cuh's streaming_context pattern — the model adapter
is the 'streaming context' that provides partition-specific information
(text_config, shared_expert, layer_types) to the generic MoE dispatch layer.

Competition: Basic award requires ALL 50+ functional tests to pass.
No one has achieved this yet. This registration is the prerequisite.
This commit is contained in:
muh-bot
2026-08-06 04:26:19 +00:00
parent d8d435c7d0
commit 08dc010a15
2 changed files with 589 additions and 0 deletions

View File

@@ -77,6 +77,7 @@ _TEXT_GENERATION_MODELS = {
"Qwen2MoeForCausalLM": ("qwen2_moe", "Qwen2MoeForCausalLM"),
"Qwen3ForCausalLM": ("qwen3", "Qwen3ForCausalLM"),
"Qwen3MoeForCausalLM": ("qwen3_moe", "Qwen3MoeForCausalLM"),
"Qwen3_5MoeForCausalLM": ("qwen3_5", "Qwen3_5MoeForCausalLM"),
"RWForCausalLM": ("falcon", "FalconForCausalLM"),
"StableLMEpochForCausalLM": ("stablelm", "StablelmForCausalLM"),
"StableLmForCausalLM": ("stablelm", "StablelmForCausalLM"),