RYDE-WORK/MiniCPM
1
0
Fork 0
mirror of https://github.com/RYDE-WORK/MiniCPM.git synced 2026-01-19 21:03:39 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

增加了自定义数据集的量化方法

Browse Source
This commit is contained in:
刘丹 2024-06-25 10:24:17 +08:00
parent 4b6474aaa6
commit 3347021a0c
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
quantize/awq_quantize.py
View File

@ -7,10 +7,11 @@ import os
model_path = '/root/ld/ld_model_pretrained/MiniCPM-1B-sft-bf16' # model_path or model_id
quant_path = '/root/ld/ld_project/pull_request/MiniCPM/quantize/awq_cpm_1b_4bit' # quant_save_path
quant_data_path='/root/ld/ld_project/pull_request/MiniCPM/quantize/quantize_data/wikitext'
quant_data_path='/root/ld/ld_project/pull_request/MiniCPM/quantize/quantize_data/wikitext'# 写入自带
quant_config = { "zero_point": True, "q_group_size": 128, "w_bit": 4, "version": "GEMM" } # "w_bit":4 or 8
quant_samples=512 # how many samples to use for calibration
custom_data=[{'question':'你叫什么名字。','answer':'我是openmbmb开源的小钢炮minicpm。'},
{'question':'你有什么特色。','answer':'我很小，但是我很强。'}]
# Load model
model = AutoAWQForCausalLM.from_pretrained(model_path)
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True,device_map={"": "cuda:0"})
@ -21,7 +22,7 @@ def load_alpaca(quant_data_path):
# concatenate data
def concatenate_data(x):
return {"text": '<s><用户>'+x['instruction'] + '<AI>' + x['input'] + '\n' + x['output']}
return {"text": '<s><用户>'+x['instruction'] + x['input'] + '<AI>' + '\n' + x['output']}
concatenated = data.map(concatenate_data)[:quant_samples]
return [text for text in concatenated["text"]]
@ -30,6 +31,9 @@ def load_wikitext(quant_data_path):
data = load_dataset(quant_data_path, split="train")
return [text for text in data["text"] if text.strip() != '' and len(text.split(' ')) > 20][:quant_samples]
def load_cust_data(custom_data):
quant_data=['<s><用户>'+i['question'] + '<AI>' + i['answer'] + '<s>' for i in custom_data]
return quant_data[:quant_samples]
# Quantize
model.quantize(tokenizer, quant_config=quant_config, calib_data=load_wikitext(quant_data_path=quant_data_path))