|
model_name = "deepseek-coder-6.7b-instruct" |
|
cmd_to_install = "ๆช็ฅ" |
|
|
|
import os |
|
from toolbox import ProxyNetworkActivate |
|
from toolbox import get_conf |
|
from .local_llm_class import LocalLLMHandle, get_local_llm_predict_fns |
|
from threading import Thread |
|
import torch |
|
|
|
def download_huggingface_model(model_name, max_retry, local_dir): |
|
from huggingface_hub import snapshot_download |
|
for i in range(1, max_retry): |
|
try: |
|
snapshot_download(repo_id=model_name, local_dir=local_dir, resume_download=True) |
|
break |
|
except Exception as e: |
|
print(f'\n\nไธ่ฝฝๅคฑ่ดฅ๏ผ้่ฏ็ฌฌ{i}ๆฌกไธญ...\n\n') |
|
return local_dir |
|
|
|
|
|
|
|
class GetCoderLMHandle(LocalLLMHandle): |
|
|
|
def load_model_info(self): |
|
|
|
self.model_name = model_name |
|
self.cmd_to_install = cmd_to_install |
|
|
|
def load_model_and_tokenizer(self): |
|
|
|
with ProxyNetworkActivate('Download_LLM'): |
|
from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer |
|
model_name = "deepseek-ai/deepseek-coder-6.7b-instruct" |
|
|
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True) |
|
self._streamer = TextIteratorStreamer(tokenizer) |
|
device_map = { |
|
"transformer.word_embeddings": 0, |
|
"transformer.word_embeddings_layernorm": 0, |
|
"lm_head": 0, |
|
"transformer.h": 0, |
|
"transformer.ln_f": 0, |
|
"model.embed_tokens": 0, |
|
"model.layers": 0, |
|
"model.norm": 0, |
|
} |
|
|
|
|
|
quantization_type = get_conf('LOCAL_MODEL_QUANT') |
|
|
|
if get_conf('LOCAL_MODEL_DEVICE') != 'cpu': |
|
if quantization_type == "INT8": |
|
from transformers import BitsAndBytesConfig |
|
|
|
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, load_in_8bit=True, |
|
device_map=device_map) |
|
elif quantization_type == "INT4": |
|
from transformers import BitsAndBytesConfig |
|
|
|
bnb_config = BitsAndBytesConfig( |
|
load_in_4bit=True, |
|
bnb_4bit_use_double_quant=True, |
|
bnb_4bit_quant_type="nf4", |
|
bnb_4bit_compute_dtype=torch.bfloat16 |
|
) |
|
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, |
|
quantization_config=bnb_config, device_map=device_map) |
|
else: |
|
|
|
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, |
|
torch_dtype=torch.bfloat16, device_map=device_map) |
|
else: |
|
|
|
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, |
|
torch_dtype=torch.bfloat16) |
|
|
|
return model, tokenizer |
|
|
|
def llm_stream_generator(self, **kwargs): |
|
|
|
def adaptor(kwargs): |
|
query = kwargs['query'] |
|
max_length = kwargs['max_length'] |
|
top_p = kwargs['top_p'] |
|
temperature = kwargs['temperature'] |
|
history = kwargs['history'] |
|
return query, max_length, top_p, temperature, history |
|
|
|
query, max_length, top_p, temperature, history = adaptor(kwargs) |
|
history.append({ 'role': 'user', 'content': query}) |
|
messages = history |
|
inputs = self._tokenizer.apply_chat_template(messages, return_tensors="pt") |
|
if inputs.shape[1] > max_length: |
|
inputs = inputs[:, -max_length:] |
|
inputs = inputs.to(self._model.device) |
|
generation_kwargs = dict( |
|
inputs=inputs, |
|
max_new_tokens=max_length, |
|
do_sample=False, |
|
top_p=top_p, |
|
streamer = self._streamer, |
|
top_k=50, |
|
temperature=temperature, |
|
num_return_sequences=1, |
|
eos_token_id=32021, |
|
) |
|
thread = Thread(target=self._model.generate, kwargs=generation_kwargs, daemon=True) |
|
thread.start() |
|
generated_text = "" |
|
for new_text in self._streamer: |
|
generated_text += new_text |
|
|
|
yield generated_text |
|
|
|
|
|
def try_to_import_special_deps(self, **kwargs): pass |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
predict_no_ui_long_connection, predict = get_local_llm_predict_fns(GetCoderLMHandle, model_name, history_format='chatglm3') |