8cb5615fb010b34c2f4f89e03e6257bfee851f86
2 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Shanshan Shen
|
bfccf739e2 |
[ModelRunner] Refactor model_runner for NPU (#6)
### What this PR does / why we need it? This PR is a refactoring of model runner, to decouple it from the classes specifically designed for GPU. The changes of model runner are generally showed below:  **Other changes:** I have removed the code of `cuda`, `lora` and `prompt adapter`, because NPU doesn`t support them now. ### Does this PR introduce _any_ user-facing change? no. ### How was this patch tested? I have used `AI-ModelScope/gpt2` for testing `examples/offline_inference_npu.py`, and the results showed that it worked well. The test logs are showed below: ```bash INFO 02-05 09:08:46 __init__.py:30] Available plugins for group vllm.platform_plugins: INFO 02-05 09:08:46 __init__.py:32] name=ascend, value=vllm_ascend:register INFO 02-05 09:08:46 __init__.py:34] all available plugins for group vllm.platform_plugins will be loaded. INFO 02-05 09:08:46 __init__.py:36] set environment variable VLLM_PLUGINS to control which plugins to load. INFO 02-05 09:08:46 __init__.py:44] plugin ascend loaded. INFO 02-05 09:08:46 __init__.py:177] Platform plugin ascend is activated INFO 02-05 09:08:48 config.py:2383] Downcasting torch.float32 to torch.float16. INFO 02-05 09:08:59 config.py:542] This model supports multiple tasks: {'generate', 'score', 'embed', 'reward', 'classify'}. Defaulting to 'generate'. INFO 02-05 09:08:59 llm_engine.py:234] Initializing a V0 LLM engine (v0.1.dev1+gb3a0d01) with config: model='/home/sss/models/AI-ModelScope/gpt2', speculative_config=None, tokenizer='/home/sss/models/AI-ModelScope/gpt2', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, override_neuron_config=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.float16, max_seq_len=1024, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=npu, decoding_config=DecodingConfig(guided_decoding_backend='xgrammar'), observability_config=ObservabilityConfig(otlp_traces_endpoint=None, collect_model_forward_time=False, collect_model_execute_time=False), seed=0, served_model_name=/home/sss/models/AI-ModelScope/gpt2, num_scheduler_steps=1, multi_step_stream_outputs=True, enable_prefix_caching=False, chunked_prefill_enabled=False, use_async_output_proc=True, disable_mm_preprocessor_cache=False, mm_processor_kwargs=None, pooler_config=None, compilation_config={"splitting_ops":[],"compile_sizes":[],"cudagraph_capture_sizes":[256,248,240,232,224,216,208,200,192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],"max_capture_size":256}, use_cached_outputs=False, WARNING 02-05 09:09:01 _custom_ops.py:21] Failed to import from vllm._C with ModuleNotFoundError("No module named 'vllm._C'") INFO 02-05 09:09:01 importing.py:16] Triton not installed or not compatible; certain GPU-related functions will not be available. Loading safetensors checkpoint shards: 0% Completed | 0/1 [00:00<?, ?it/s] Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:00<00:00, 3.18it/s] Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:00<00:00, 3.18it/s] INFO 02-05 09:09:11 executor_base.py:110] # CPU blocks: 98557, # CPU blocks: 7281 INFO 02-05 09:09:11 executor_base.py:115] Maximum concurrency for 1024 tokens per request: 1539.95x INFO 02-05 09:09:12 llm_engine.py:431] init engine (profile, create kv cache, warmup model) took 2.13 seconds Processed prompts: 100%|██████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:02<00:00, 1.53it/s, est. speed input: 8.41 toks/s, output: 152.97 toks/s] Prompt: 'Hello, my name is', Generated text: " John. I'm a writer, and I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm a writer. I'm" Prompt: 'The president of the United States is', Generated text: ' States president. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United States. He is the president of the United' Prompt: 'The capital of France is', Generated text: ' the capital of the French Republic, and the capital of the French Republic is the capital of the French Republic.\n\nThe French Republic is the capital of the French Republic.\n\nThe French Republic is the capital of the French Republic.\n\nThe French Republic is the capital of the French Republic.\n\nThe French Republic is the capital of the French Republic.\n\nThe French Republic is the capital of the French Republic.\n\nThe French Republic is the capital of the French Republic.' Prompt: 'The future of AI is', Generated text: '\n\nThe future of AI is a question of how to make it work.\n\nThe future of AI is a question of how to make it work.\n\nThe future of AI is a question of how to make it work.\n\nThe future of AI is a question of how to make it work.\n\nThe future of AI is a question of how to make it work.\n\nThe future of AI is a question of how to make it work.\n\nThe future' ``` --------- Signed-off-by: Shanshan Shen <467638484@qq.com> |
||
|
Yikun Jiang
|
d5e7756028 |
[Core] Init vllm-ascend (#3)
### What this PR does / why we need it?
vLLM Ascend plugin (vllm-ascend) is a backend plugin for running vLLM on
the Ascend NPU.
This plugin is the recommended approach for supporting the Ascend
backend within the vLLM community. It adheres to the principles outlined
in the [RFC]: Hardware pluggable, providing a hardware-pluggable
interface that decouples the integration of the Ascend NPU with vLLM.
This patch also include changes to make CI work and use cache speed up
e2e test, including:
1. Change push (post merge ci) and pull_request (pr ci) trigger branch
to main
2. Make mypy work by ignore base_communicator and clear unused deps
3. Several improvements for vllm_ascend_test:
- use cache (pip, ms, hf) speed up e2e test (25mins --> 5mins)
- switch `git clone` command to `action/checkout` to speedup checkout
and
- Enable sv for pytest for better info dump
- Remove network host to resole `docker: conflicting ontions: cannot
attach both user-defined and non-user-definednetwork-modes`, which is a
problem on docker 1.45 but not on 1.39.
4. Adapt MLA decode optimizations:
|