deepgeek/MiniCPM-o
2
0
Fork 0
mirror of https://github.com/OpenBMB/MiniCPM-V.git synced 2026-02-05 18:29:18 +08:00
Code Issues Packages Projects Releases Wiki Activity

修改了代码，增加了注释和规范性

Browse Source
This commit is contained in:
root
2024-07-12 16:09:49 +08:00
parent bee803ffaa
commit e47c6ab5aa
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
quantize/bnb_quantize.py
View File

@@ -40,12 +40,12 @@ quantization_config = BitsAndBytesConfig(
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True) tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
model = AutoModel.from_pretrained( model = AutoModel.from_pretrained(
model_path, model_path,
device_map="cuda:0", # 分配模型到GPU0 device_map=device, # 分配模型到device
quantization_config=quantization_config, quantization_config=quantization_config,
trust_remote_code=True trust_remote_code=True
) )
gpu_usage = GPUtil.getGPUs()[0].memoryUsed
gpu_usage = GPUtil.getGPUs()[0].memoryUsed
start=time.time() start=time.time()
response = model.chat( response = model.chat(
image=Image.open(image_path).convert("RGB"), image=Image.open(image_path).convert("RGB"),
@@ -58,7 +58,7 @@ response = model.chat(
tokenizer=tokenizer tokenizer=tokenizer
) # 模型推理 ) # 模型推理
print('量化后输出',response) print('量化后输出',response)
print('量化后用时',time.time()-start) print('量化后推理用时',time.time()-start)
print(f"量化后显存占用: {round(gpu_usage/1024,2)}GB") print(f"量化后显存占用: {round(gpu_usage/1024,2)}GB")
""" """