OsakanaTeishoku commited on
Commit
a3cd682
1 Parent(s): 74785a1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -12
README.md CHANGED
@@ -48,15 +48,7 @@ from jinja2 import Template
48
  print(vllm.__version__)
49
 
50
  MAX_LENGTH = 1024
51
- # BATCH_SIZE = 2
52
- # MODEL_NAME = "/kaggle/input/gemma/transformers/2b-it/3"
53
- # CHECKPOINT_PATH = "/kaggle/input/atmacup17-train-causal/checkpoint-468"
54
- #MODEL_NAME = "/kaggle/input/gemma2-9b-it-for-t4"
55
- #CHECKPOINT_PATH = "/kaggle/input/exp002-gemma-2-9b-it"
56
- #MODEL_NAME = "team-hatakeyama-phase2/Tanuki-8B-dpo-v1.0-AWQ"
57
- #MODEL_NAME = "llm-jp/llm-jp-3-3.7b-instruct"
58
  MODEL_NAME = "llm-jp/llm-jp-3-13b"
59
- #MODEL_NAME = "OsakanaTeishoku/1127-13b-magpie-sft-step170"
60
  print(MODEL_NAME)
61
 
62
  import os
@@ -64,16 +56,13 @@ os.environ["HF_TOKEN"] = "あなたのHugging Faceトークン"
64
 
65
  from vllm.lora.request import LoRARequest
66
  llm = vllm.LLM(
67
- MODEL_NAME, # "deepseek-ai/deepseek-math-7b-instruct"
68
  tensor_parallel_size=1, # 2, 4
69
  gpu_memory_utilization=0.95,
70
  trust_remote_code=True,
71
- #enable_lora=True,
72
- #dtype="half",
73
  enforce_eager=True,
74
  max_model_len=MAX_LENGTH,
75
  enable_lora=True,
76
- #quantization="AWQ",
77
  quantization="bitsandbytes",
78
  load_format="bitsandbytes"
79
  )
 
48
  print(vllm.__version__)
49
 
50
  MAX_LENGTH = 1024
 
 
 
 
 
 
 
51
  MODEL_NAME = "llm-jp/llm-jp-3-13b"
 
52
  print(MODEL_NAME)
53
 
54
  import os
 
56
 
57
  from vllm.lora.request import LoRARequest
58
  llm = vllm.LLM(
59
+ MODEL_NAME,
60
  tensor_parallel_size=1, # 2, 4
61
  gpu_memory_utilization=0.95,
62
  trust_remote_code=True,
 
 
63
  enforce_eager=True,
64
  max_model_len=MAX_LENGTH,
65
  enable_lora=True,
 
66
  quantization="bitsandbytes",
67
  load_format="bitsandbytes"
68
  )