初始化项目，由ModelHub XC社区提供模型
Model: NotoriousH2/gemma-3-1b-it-Math-SFT Source: Original Platform
This commit is contained in:
ModelHub XC
36
.gitattributes
vendored
Normal file
36
.gitattributes
vendored
Normal file
@@ -0,0 +1,36 @@
|
||||
*.7z filter=lfs diff=lfs merge=lfs -text
|
||||
*.arrow filter=lfs diff=lfs merge=lfs -text
|
||||
*.bin filter=lfs diff=lfs merge=lfs -text
|
||||
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
||||
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
||||
*.ftz filter=lfs diff=lfs merge=lfs -text
|
||||
*.gz filter=lfs diff=lfs merge=lfs -text
|
||||
*.h5 filter=lfs diff=lfs merge=lfs -text
|
||||
*.joblib filter=lfs diff=lfs merge=lfs -text
|
||||
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
||||
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
||||
*.model filter=lfs diff=lfs merge=lfs -text
|
||||
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
||||
*.npy filter=lfs diff=lfs merge=lfs -text
|
||||
*.npz filter=lfs diff=lfs merge=lfs -text
|
||||
*.onnx filter=lfs diff=lfs merge=lfs -text
|
||||
*.ot filter=lfs diff=lfs merge=lfs -text
|
||||
*.parquet filter=lfs diff=lfs merge=lfs -text
|
||||
*.pb filter=lfs diff=lfs merge=lfs -text
|
||||
*.pickle filter=lfs diff=lfs merge=lfs -text
|
||||
*.pkl filter=lfs diff=lfs merge=lfs -text
|
||||
*.pt filter=lfs diff=lfs merge=lfs -text
|
||||
*.pth filter=lfs diff=lfs merge=lfs -text
|
||||
*.rar filter=lfs diff=lfs merge=lfs -text
|
||||
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
||||
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
||||
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
||||
*.tar filter=lfs diff=lfs merge=lfs -text
|
||||
*.tflite filter=lfs diff=lfs merge=lfs -text
|
||||
*.tgz filter=lfs diff=lfs merge=lfs -text
|
||||
*.wasm filter=lfs diff=lfs merge=lfs -text
|
||||
*.xz filter=lfs diff=lfs merge=lfs -text
|
||||
*.zip filter=lfs diff=lfs merge=lfs -text
|
||||
*.zst filter=lfs diff=lfs merge=lfs -text
|
||||
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
||||
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
||||
140
README.md
Normal file
140
README.md
Normal file
@@ -0,0 +1,140 @@
|
||||
---
|
||||
language: ko
|
||||
license: apache-2.0
|
||||
base_model: google/gemma-3-1b-it
|
||||
tags:
|
||||
- math
|
||||
- korean
|
||||
- sft
|
||||
- gemma
|
||||
- distillation
|
||||
datasets:
|
||||
- NotoriousH2/HRM8K
|
||||
---
|
||||
|
||||
# Gemma-3-1B-IT Math SFT
|
||||
|
||||
`google/gemma-3-1b-it`를 한국어 수학 문제(GSM8K)에 대해 교사 증류 SFT한 모델.
|
||||
|
||||
## 성능
|
||||
|
||||
| Benchmark | Score |
|
||||
|-----------|-------|
|
||||
| HRM8K eval GSM8K (264문제, Korean) | **~44.9%** (3회 평균) |
|
||||
| HRM8K eval MATH (577문제, Korean) | ~17% |
|
||||
|
||||
평가: temperature=0, vLLM 서빙, max_tokens=2048
|
||||
|
||||
## 데이터 생성 파이프라인
|
||||
|
||||
### 원본 데이터
|
||||
- **GSM8K train set**: 영어 초등 수학 7,473문제 ([openai/gsm8k](https://huggingface.co/datasets/openai/gsm8k))
|
||||
- **평가 데이터**: [HRM8K](https://huggingface.co/datasets/NotoriousH2/HRM8K) eval set 841문제 (GSM8K 264 + MATH 577, 한국어)
|
||||
|
||||
### 교사 모델
|
||||
- **모델**: (AWQ 4bit)
|
||||
- **서빙**: vLLM 0.11.0 (구버전 필요, AWQ 호환),
|
||||
- **중요**: 이 교사 모델은 HRM8K 학습 데이터를 생성한 바로 그 모델. 다른 교사 모델(Qwen3.5-9B/35B 등)은 스타일 불일치로 -10%p 이상 성능 하락.
|
||||
|
||||
### 2단계 데이터 생성
|
||||
|
||||
|
||||
### 최종 학습 데이터 형식
|
||||
|
||||
- 총 26,254개 (train 95% / eval 5% split)
|
||||
- 시스템 프롬프트:
|
||||
|
||||
## 학습 설정
|
||||
|
||||
|
||||
|
||||
## 재현 방법
|
||||
|
||||
INFO 03-19 14:51:58 [__init__.py:216] Automatically detected platform cuda.
|
||||
[1;36m(APIServer pid=3426235)[0;0m INFO 03-19 14:52:04 [api_server.py:1839] vLLM API server version 0.11.0
|
||||
[1;36m(APIServer pid=3426235)[0;0m INFO 03-19 14:52:04 [utils.py:233] non-default args: {'model_tag': 'cpatonn/Qwen3-30B-A3B-Instruct-2507-AWQ-4bit', 'port': 8001, 'model': 'cpatonn/Qwen3-30B-A3B-Instruct-2507-AWQ-4bit', 'max_model_len': 4096, 'gpu_memory_utilization': 0.8}
|
||||
[1;36m(APIServer pid=3426235)[0;0m INFO 03-19 14:52:06 [model.py:547] Resolved architecture: Qwen3MoeForCausalLM
|
||||
[1;36m(APIServer pid=3426235)[0;0m INFO 03-19 14:52:06 [model.py:1510] Using max model len 4096
|
||||
[1;36m(APIServer pid=3426235)[0;0m INFO 03-19 14:52:07 [scheduler.py:205] Chunked prefill is enabled with max_num_batched_tokens=8192.
|
||||
INFO 03-19 14:52:19 [__init__.py:216] Automatically detected platform cuda.
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m INFO 03-19 14:52:25 [core.py:644] Waiting for init message from front-end.
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m INFO 03-19 14:52:25 [core.py:77] Initializing a V1 LLM engine (v0.11.0) with config: model='cpatonn/Qwen3-30B-A3B-Instruct-2507-AWQ-4bit', speculative_config=None, tokenizer='cpatonn/Qwen3-30B-A3B-Instruct-2507-AWQ-4bit', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=4096, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, data_parallel_size=1, disable_custom_all_reduce=False, quantization=compressed-tensors, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend='auto', disable_fallback=False, disable_any_whitespace=False, disable_additional_properties=False, reasoning_parser=''), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None), seed=0, served_model_name=cpatonn/Qwen3-30B-A3B-Instruct-2507-AWQ-4bit, enable_prefix_caching=True, chunked_prefill_enabled=True, pooler_config=None, compilation_config={"level":3,"debug_dump_path":"","cache_dir":"","backend":"","custom_ops":[],"splitting_ops":["vllm.unified_attention","vllm.unified_attention_with_output","vllm.mamba_mixer2","vllm.mamba_mixer","vllm.short_conv","vllm.linear_attention","vllm.plamo2_mamba_mixer","vllm.gdn_attention","vllm.sparse_attn_indexer"],"use_inductor":true,"compile_sizes":[],"inductor_compile_config":{"enable_auto_functionalized_v2":false},"inductor_passes":{},"cudagraph_mode":[2,1],"use_cudagraph":true,"cudagraph_num_of_warmups":1,"cudagraph_capture_sizes":[512,504,496,488,480,472,464,456,448,440,432,424,416,408,400,392,384,376,368,360,352,344,336,328,320,312,304,296,288,280,272,264,256,248,240,232,224,216,208,200,192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],"cudagraph_copy_inputs":false,"full_cuda_graph":false,"use_inductor_graph_partition":false,"pass_config":{},"max_capture_size":512,"local_cache_dir":null}
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m INFO 03-19 14:52:26 [parallel_state.py:1208] rank 0 in world size 1 is assigned as DP rank 0, PP rank 0, TP rank 0, EP rank 0
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] EngineCore failed to start.
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] Traceback (most recent call last):
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/engine/core.py", line 699, in run_engine_core
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] engine_core = EngineCoreProc(*args, **kwargs)
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/engine/core.py", line 498, in __init__
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] super().__init__(vllm_config, executor_class, log_stats,
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/engine/core.py", line 83, in __init__
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] self.model_executor = executor_class(vllm_config)
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/executor/executor_base.py", line 54, in __init__
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] self._init_executor()
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/executor/uniproc_executor.py", line 54, in _init_executor
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] self.collective_rpc("init_device")
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/executor/uniproc_executor.py", line 83, in collective_rpc
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] return [run_method(self.driver_worker, method, args, kwargs)]
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/utils/__init__.py", line 3122, in run_method
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] return func(*args, **kwargs)
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/worker/worker_base.py", line 259, in init_device
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] self.worker.init_device() # type: ignore
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/worker/gpu_worker.py", line 187, in init_device
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] raise ValueError(
|
||||
[1;36m(EngineCore_DP0 pid=3426796)[0;0m ERROR 03-19 14:52:26 [core.py:708] ValueError: Free memory on device (7.29/93.1 GiB) on startup is less than desired GPU memory utilization (0.8, 74.48 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.
|
||||
INFO 03-19 14:52:34 [__init__.py:216] Automatically detected platform cuda.
|
||||
[1;36m(APIServer pid=3427292)[0;0m INFO 03-19 14:52:40 [api_server.py:1839] vLLM API server version 0.11.0
|
||||
[1;36m(APIServer pid=3427292)[0;0m INFO 03-19 14:52:40 [utils.py:233] non-default args: {'model_tag': './outputs/models/c17d-gemma-3-1b-it-Math', 'model': './outputs/models/c17d-gemma-3-1b-it-Math', 'dtype': 'bfloat16', 'max_model_len': 4096, 'gpu_memory_utilization': 0.85}
|
||||
[1;36m(APIServer pid=3427292)[0;0m INFO 03-19 14:52:51 [model.py:547] Resolved architecture: Gemma3ForCausalLM
|
||||
[1;36m(APIServer pid=3427292)[0;0m INFO 03-19 14:52:51 [model.py:1510] Using max model len 4096
|
||||
[1;36m(APIServer pid=3427292)[0;0m INFO 03-19 14:52:51 [scheduler.py:205] Chunked prefill is enabled with max_num_batched_tokens=8192.
|
||||
INFO 03-19 14:52:58 [__init__.py:216] Automatically detected platform cuda.
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m INFO 03-19 14:53:03 [core.py:644] Waiting for init message from front-end.
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m INFO 03-19 14:53:03 [core.py:77] Initializing a V1 LLM engine (v0.11.0) with config: model='./outputs/models/c17d-gemma-3-1b-it-Math', speculative_config=None, tokenizer='./outputs/models/c17d-gemma-3-1b-it-Math', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=4096, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, data_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend='auto', disable_fallback=False, disable_any_whitespace=False, disable_additional_properties=False, reasoning_parser=''), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None), seed=0, served_model_name=./outputs/models/c17d-gemma-3-1b-it-Math, enable_prefix_caching=True, chunked_prefill_enabled=True, pooler_config=None, compilation_config={"level":3,"debug_dump_path":"","cache_dir":"","backend":"","custom_ops":[],"splitting_ops":["vllm.unified_attention","vllm.unified_attention_with_output","vllm.mamba_mixer2","vllm.mamba_mixer","vllm.short_conv","vllm.linear_attention","vllm.plamo2_mamba_mixer","vllm.gdn_attention","vllm.sparse_attn_indexer"],"use_inductor":true,"compile_sizes":[],"inductor_compile_config":{"enable_auto_functionalized_v2":false},"inductor_passes":{},"cudagraph_mode":[2,1],"use_cudagraph":true,"cudagraph_num_of_warmups":1,"cudagraph_capture_sizes":[512,504,496,488,480,472,464,456,448,440,432,424,416,408,400,392,384,376,368,360,352,344,336,328,320,312,304,296,288,280,272,264,256,248,240,232,224,216,208,200,192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],"cudagraph_copy_inputs":false,"full_cuda_graph":false,"use_inductor_graph_partition":false,"pass_config":{},"max_capture_size":512,"local_cache_dir":null}
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m INFO 03-19 14:53:05 [parallel_state.py:1208] rank 0 in world size 1 is assigned as DP rank 0, PP rank 0, TP rank 0, EP rank 0
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] EngineCore failed to start.
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] Traceback (most recent call last):
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/engine/core.py", line 699, in run_engine_core
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] engine_core = EngineCoreProc(*args, **kwargs)
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/engine/core.py", line 498, in __init__
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] super().__init__(vllm_config, executor_class, log_stats,
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/engine/core.py", line 83, in __init__
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] self.model_executor = executor_class(vllm_config)
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/executor/executor_base.py", line 54, in __init__
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] self._init_executor()
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/executor/uniproc_executor.py", line 54, in _init_executor
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] self.collective_rpc("init_device")
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/executor/uniproc_executor.py", line 83, in collective_rpc
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] return [run_method(self.driver_worker, method, args, kwargs)]
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/utils/__init__.py", line 3122, in run_method
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] return func(*args, **kwargs)
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/worker/worker_base.py", line 259, in init_device
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] self.worker.init_device() # type: ignore
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] File "/tmp/.venv/lib/python3.12/site-packages/vllm/v1/worker/gpu_worker.py", line 187, in init_device
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] raise ValueError(
|
||||
[1;36m(EngineCore_DP0 pid=3428117)[0;0m ERROR 03-19 14:53:05 [core.py:708] ValueError: Free memory on device (7.29/93.1 GiB) on startup is less than desired GPU memory utilization (0.85, 79.13 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.
|
||||
|
||||
## 파일
|
||||
- : SFT 학습 스크립트
|
||||
- : HRM8K 평가 스크립트 (vLLM OpenAI API 호환 서버 필요)
|
||||
3
added_tokens.json
Normal file
3
added_tokens.json
Normal file
@@ -0,0 +1,3 @@
|
||||
{
|
||||
"<image_soft_token>": 262144
|
||||
}
|
||||
47
chat_template.jinja
Normal file
47
chat_template.jinja
Normal file
@@ -0,0 +1,47 @@
|
||||
{{ bos_token }}
|
||||
{%- if messages[0]['role'] == 'system' -%}
|
||||
{%- if messages[0]['content'] is string -%}
|
||||
{%- set first_user_prefix = messages[0]['content'] + '
|
||||
|
||||
' -%}
|
||||
{%- else -%}
|
||||
{%- set first_user_prefix = messages[0]['content'][0]['text'] + '
|
||||
|
||||
' -%}
|
||||
{%- endif -%}
|
||||
{%- set loop_messages = messages[1:] -%}
|
||||
{%- else -%}
|
||||
{%- set first_user_prefix = "" -%}
|
||||
{%- set loop_messages = messages -%}
|
||||
{%- endif -%}
|
||||
{%- for message in loop_messages -%}
|
||||
{%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
|
||||
{{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
|
||||
{%- endif -%}
|
||||
{%- if (message['role'] == 'assistant') -%}
|
||||
{%- set role = "model" -%}
|
||||
{%- else -%}
|
||||
{%- set role = message['role'] -%}
|
||||
{%- endif -%}
|
||||
{{ '<start_of_turn>' + role + '
|
||||
' + (first_user_prefix if loop.first else "") }}
|
||||
{%- if message['content'] is string -%}
|
||||
{{ message['content'] | trim }}
|
||||
{%- elif message['content'] is iterable -%}
|
||||
{%- for item in message['content'] -%}
|
||||
{%- if item['type'] == 'image' -%}
|
||||
{{ '<start_of_image>' }}
|
||||
{%- elif item['type'] == 'text' -%}
|
||||
{{ item['text'] | trim }}
|
||||
{%- endif -%}
|
||||
{%- endfor -%}
|
||||
{%- else -%}
|
||||
{{ raise_exception("Invalid content type") }}
|
||||
{%- endif -%}
|
||||
{{ '<end_of_turn>
|
||||
' }}
|
||||
{%- endfor -%}
|
||||
{%- if add_generation_prompt -%}
|
||||
{{'<start_of_turn>model
|
||||
'}}
|
||||
{%- endif -%}
|
||||
64
config.json
Normal file
64
config.json
Normal file
@@ -0,0 +1,64 @@
|
||||
{
|
||||
"_sliding_window_pattern": 6,
|
||||
"architectures": [
|
||||
"Gemma3ForCausalLM"
|
||||
],
|
||||
"attention_bias": false,
|
||||
"attention_dropout": 0.0,
|
||||
"attn_logit_softcapping": null,
|
||||
"bos_token_id": 2,
|
||||
"cache_implementation": "hybrid",
|
||||
"dtype": "bfloat16",
|
||||
"eos_token_id": 1,
|
||||
"final_logit_softcapping": null,
|
||||
"head_dim": 256,
|
||||
"hidden_activation": "gelu_pytorch_tanh",
|
||||
"hidden_size": 1152,
|
||||
"initializer_range": 0.02,
|
||||
"intermediate_size": 6912,
|
||||
"layer_types": [
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"full_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"full_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"full_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention",
|
||||
"full_attention",
|
||||
"sliding_attention",
|
||||
"sliding_attention"
|
||||
],
|
||||
"max_position_embeddings": 32768,
|
||||
"model_type": "gemma3_text",
|
||||
"num_attention_heads": 4,
|
||||
"num_hidden_layers": 26,
|
||||
"num_key_value_heads": 1,
|
||||
"pad_token_id": 1,
|
||||
"query_pre_attn_scalar": 256,
|
||||
"rms_norm_eps": 1e-06,
|
||||
"rope_local_base_freq": 10000,
|
||||
"rope_scaling": null,
|
||||
"rope_theta": 1000000,
|
||||
"sliding_window": 512,
|
||||
"sliding_window_pattern": 6,
|
||||
"transformers_version": "4.57.3",
|
||||
"use_bidirectional_attention": false,
|
||||
"use_cache": false,
|
||||
"vocab_size": 262144
|
||||
}
|
||||
86
eval.py
Normal file
86
eval.py
Normal file
@@ -0,0 +1,86 @@
|
||||
"""공통 평가 스크립트: vLLM 서버에 연결하여 HRM8K 전체 841문제 평가 (temperature=0)"""
|
||||
import os, json, re, sys, asyncio
|
||||
from langchain_openai import ChatOpenAI
|
||||
from langchain_core.prompts import ChatPromptTemplate
|
||||
from langchain_core.output_parsers import StrOutputParser
|
||||
from openai import OpenAI
|
||||
|
||||
MATH_SYSTEM_PROMPT = """주어진 수학 문제를 단계별로 풀고 답변을 작성하세요.
|
||||
반드시 최종 답변을 \\boxed{정수} 형식으로 마지막 줄에 출력하세요.
|
||||
예시: \\boxed{42}"""
|
||||
|
||||
def extract_boxed(text):
|
||||
m = re.findall(r'\\boxed\{([^}]+)\}', text)
|
||||
return m[-1].strip() if m else None
|
||||
|
||||
def normalize(a):
|
||||
if a is None: return None
|
||||
s = str(a).replace(",","").replace(" ","").strip()
|
||||
try:
|
||||
n = float(s)
|
||||
return str(int(n)) if n == int(n) else str(n)
|
||||
except: return s
|
||||
|
||||
def check(pred, gt):
|
||||
p, g = normalize(pred), normalize(gt)
|
||||
return p is not None and g is not None and p == g
|
||||
|
||||
async def evaluate(label="", save_path=None):
|
||||
client = OpenAI(base_url="http://localhost:8000/v1", api_key="token-abc123")
|
||||
model_name = client.models.list().data[0].id
|
||||
print(f"모델: {model_name}")
|
||||
|
||||
with open("data/HRM8k_eval.json") as f:
|
||||
data = json.load(f)
|
||||
print(f"평가: {len(data)}개 (temperature=0, max_tokens=2048)")
|
||||
|
||||
llm = ChatOpenAI(base_url="http://localhost:8000/v1", api_key="token-abc123",
|
||||
model=model_name, temperature=0, max_tokens=2048)
|
||||
prompt = ChatPromptTemplate([("user", "{sp}\n\n{q}")]).partial(sp=MATH_SYSTEM_PROMPT)
|
||||
chain = prompt | llm | StrOutputParser()
|
||||
inputs = [{"q": item["question"]} for item in data]
|
||||
results = await chain.abatch(inputs, config={"max_concurrency": 400})
|
||||
|
||||
by_src = {}
|
||||
details = []
|
||||
for item, res in zip(data, results):
|
||||
s = item.get("source", "?")
|
||||
if s not in by_src: by_src[s] = {"correct": 0, "total": 0, "no_boxed": 0}
|
||||
by_src[s]["total"] += 1
|
||||
pred = extract_boxed(res)
|
||||
is_correct = False
|
||||
if pred is None:
|
||||
by_src[s]["no_boxed"] += 1
|
||||
elif check(pred, item["answer"]):
|
||||
by_src[s]["correct"] += 1
|
||||
is_correct = True
|
||||
details.append({
|
||||
"question": item["question"][:80],
|
||||
"source": s,
|
||||
"gt": str(item["answer"])[-30:] if isinstance(item["answer"], str) else str(item["answer"]),
|
||||
"pred": pred,
|
||||
"correct": is_correct,
|
||||
})
|
||||
|
||||
tc = sum(v["correct"] for v in by_src.values())
|
||||
tt = sum(v["total"] for v in by_src.values())
|
||||
print(f"\n=== {label} 결과 (temperature=0) ===")
|
||||
for s in sorted(by_src):
|
||||
v = by_src[s]
|
||||
print(f" [{s.upper()}] {v['correct']}/{v['total']} ({v['correct']/v['total']*100:.1f}%) | boxed미출력: {v['no_boxed']}")
|
||||
print(f" [전체] {tc}/{tt} ({tc/tt*100:.1f}%)")
|
||||
|
||||
result_obj = {"label": label, "correct": tc, "total": tt, "accuracy": tc/tt*100, "by_source": by_src}
|
||||
|
||||
if save_path:
|
||||
os.makedirs(os.path.dirname(save_path) or ".", exist_ok=True)
|
||||
with open(save_path, "w") as f:
|
||||
json.dump({"result": result_obj, "details": details}, f, ensure_ascii=False, indent=2)
|
||||
print(f" 결과 저장: {save_path}")
|
||||
|
||||
return result_obj
|
||||
|
||||
if __name__ == "__main__":
|
||||
label = sys.argv[1] if len(sys.argv) > 1 else "eval"
|
||||
save_path = sys.argv[2] if len(sys.argv) > 2 else None
|
||||
asyncio.run(evaluate(label, save_path))
|
||||
14
generation_config.json
Normal file
14
generation_config.json
Normal file
@@ -0,0 +1,14 @@
|
||||
{
|
||||
"bos_token_id": 2,
|
||||
"cache_implementation": "hybrid",
|
||||
"do_sample": true,
|
||||
"eos_token_id": [
|
||||
1,
|
||||
1,
|
||||
106
|
||||
],
|
||||
"pad_token_id": 1,
|
||||
"top_k": 64,
|
||||
"top_p": 0.95,
|
||||
"transformers_version": "4.57.3"
|
||||
}
|
||||
3
pytorch_model.bin
Normal file
3
pytorch_model.bin
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:2f93d8e31ba1ce5c611b784c4cba1082eef27560a277c0537da1dfe59db6c498
|
||||
size 1999887347
|
||||
27
special_tokens_map.json
Normal file
27
special_tokens_map.json
Normal file
@@ -0,0 +1,27 @@
|
||||
{
|
||||
"boi_token": "<start_of_image>",
|
||||
"bos_token": {
|
||||
"content": "<bos>",
|
||||
"lstrip": false,
|
||||
"normalized": false,
|
||||
"rstrip": false,
|
||||
"single_word": false
|
||||
},
|
||||
"eoi_token": "<end_of_image>",
|
||||
"eos_token": {
|
||||
"content": "<eos>",
|
||||
"lstrip": false,
|
||||
"normalized": false,
|
||||
"rstrip": false,
|
||||
"single_word": false
|
||||
},
|
||||
"image_token": "<image_soft_token>",
|
||||
"pad_token": "<eos>",
|
||||
"unk_token": {
|
||||
"content": "<unk>",
|
||||
"lstrip": false,
|
||||
"normalized": false,
|
||||
"rstrip": false,
|
||||
"single_word": false
|
||||
}
|
||||
}
|
||||
3
tokenizer.json
Normal file
3
tokenizer.json
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
|
||||
size 33384568
|
||||
3
tokenizer.model
Normal file
3
tokenizer.model
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
|
||||
size 4689074
|
||||
51345
tokenizer_config.json
Normal file
51345
tokenizer_config.json
Normal file
File diff suppressed because it is too large
Load Diff
67
train_sft.py
Normal file
67
train_sft.py
Normal file
@@ -0,0 +1,67 @@
|
||||
"""C17d: 모든 풀이 + 길이 필터 (1500자 이하만) + NaN 방지"""
|
||||
import json, re, random, torch, numpy as np, os
|
||||
from collections import defaultdict
|
||||
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||||
from trl import SFTTrainer, SFTConfig
|
||||
from transformers import EarlyStoppingCallback
|
||||
from datasets import Dataset
|
||||
|
||||
SEED = 42
|
||||
random.seed(SEED); np.random.seed(SEED); torch.manual_seed(SEED)
|
||||
if torch.cuda.is_available():
|
||||
torch.cuda.manual_seed_all(SEED)
|
||||
if torch.cuda.get_device_capability()[0] >= 8: torch.set_float32_matmul_precision('high')
|
||||
|
||||
SP = "주어진 수학 문제를 단계별로 풀고 답변을 작성하세요.\n반드시 최종 답변을 \\boxed{정수} 형식으로 마지막 줄에 출력하세요.\n예시: \\boxed{42}"
|
||||
|
||||
print("=== C17d: All solutions, length-filtered (≤1500 chars) ===")
|
||||
|
||||
with open("data/GSM8K_full_qwen3_30b.json") as f:
|
||||
data = json.load(f)
|
||||
|
||||
# 길이 필터: 1500자 이하만
|
||||
filtered = [d for d in data if len(d['answer']) <= 1500]
|
||||
print(f"원본: {len(data)}개 → 필터 후: {len(filtered)}개 (제거: {len(data)-len(filtered)})")
|
||||
|
||||
random.shuffle(filtered)
|
||||
uq = len(set(d["question"] for d in filtered))
|
||||
print(f"Unique: {uq}, avg {len(filtered)/uq:.1f}/q")
|
||||
|
||||
split = int(len(filtered) * 0.95)
|
||||
train, test = filtered[:split], filtered[split:]
|
||||
def to_sft(ex):
|
||||
return {"prompt": [{"role":"user","content":SP+"\n\n"+ex["question"]}],
|
||||
"completion": [{"role":"assistant","content":ex["answer"]}]}
|
||||
|
||||
cols = [c for c in Dataset.from_list(train[:1]).column_names if c not in ["prompt","completion"]]
|
||||
train_ds = Dataset.from_list(train).map(to_sft, remove_columns=cols)
|
||||
test_ds = Dataset.from_list(test).map(to_sft, remove_columns=cols)
|
||||
print(f"학습: {len(train_ds)} / 검증: {len(test_ds)}")
|
||||
|
||||
tokenizer = AutoTokenizer.from_pretrained("outputs/models/gemma-3-1b-it")
|
||||
model = AutoModelForCausalLM.from_pretrained("outputs/models/gemma-3-1b-it", dtype=torch.bfloat16, device_map="auto", attn_implementation='flash_attention_2')
|
||||
tokenizer.pad_token = tokenizer.eos_token
|
||||
model.gradient_checkpointing_enable(); model.config.use_cache = False
|
||||
|
||||
cfg = SFTConfig(
|
||||
report_to='none', seed=SEED, eval_strategy="steps", eval_steps=200,
|
||||
save_total_limit=2, load_best_model_at_end=True, metric_for_best_model="eval_loss",
|
||||
save_steps=200, num_train_epochs=3, warmup_ratio=0.05, weight_decay=0.01, max_grad_norm=1.0,
|
||||
neftune_noise_alpha=5, per_device_train_batch_size=8, gradient_accumulation_steps=4,
|
||||
per_device_eval_batch_size=2, max_length=2048, lr_scheduler_type='cosine',
|
||||
learning_rate=2e-5, bf16=True, optim="paged_adamw_8bit",
|
||||
output_dir="outputs/c17d_checkpoints", logging_steps=50, save_strategy="steps",
|
||||
)
|
||||
|
||||
trainer = SFTTrainer(model=model, processing_class=tokenizer, train_dataset=train_ds, eval_dataset=test_ds, args=cfg,
|
||||
callbacks=[EarlyStoppingCallback(early_stopping_patience=3)])
|
||||
print("학습 시작 (3 epochs, 모든 풀이, ≤1500자)")
|
||||
r = trainer.train()
|
||||
print(f"완료! Loss: {r.training_loss:.4f}")
|
||||
|
||||
SAVE = "outputs/models/c17d-gemma-3-1b-it-Math"
|
||||
os.makedirs(SAVE, exist_ok=True)
|
||||
model.eval(); model.save_pretrained(SAVE, safe_serialization=False); tokenizer.save_pretrained(SAVE)
|
||||
print(f"저장: {SAVE}")
|
||||
del model, trainer; torch.cuda.empty_cache()
|
||||
print("GPU 해제")
|
||||
Reference in New Issue
Block a user