Large model download: Interlink Hi-Tech
ClueAI/PromptCLUE-base-v1-5 at main (huggingface.co) supports multi-task generation, supports Chinese, does not support multiple rounds of dialogue, experience: ClueAI (cluebenchmarks.com)
A further trained model based on promptclue-base: ClueAI/ChatYuan-large-v1 at main (huggingface.co) supports multi-task generation, supports Chinese, and supports simple dialogue
About huggingface model download:
Manual download: https://mirrors.tuna.tsinghua.edu.cn/hugging-face-models/hfl/
Code download:
import llama
#MODEL = '/home/guo/llama_test/llama_model'
MODEL = 'decapoda-research/llama-7b-hf'
# MODEL = 'decapoda-research/llama-13b-hf'
# MODEL = 'decapoda-research/llama-30b-hf'
# MODEL = 'decapoda-research/llama-65b-hf'
#tokenizer = llama.LLaMATokenizer.from_pretrained(MODEL,mirror='tuna')
#model = llama.LLaMAForCausalLM.from_pretrained(MODEL, mirror='tuna',low_cpu_mem_usage = True)
tokenizer = llama.LLaMATokenizer.from_pretrained(MODEL,mirror='https://mirrors.tuna.tsinghua.edu.cn/hugging-face-models')
model = llama.LLaMAForCausalLM.from_pretrained(MODEL, mirror='https://mirrors.tuna.tsinghua.edu.cn/hugging-face-models',low_cpu_mem_usage = True)
model.to('cpu')
batch = tokenizer("Yo mama", return_tensors = "pt")
print(tokenizer.decode(model.generate(batch["input_ids"], max_length=100)[0]))Regarding the download of github's mirror warehouse, refer to: (4 messages) git clone source change/GitHub domestic mirror_git change source_Mianli Duojiatang blog-CSDN blog :
https://gitclone.com
# 服务器位于杭州（可用）
使用方式：原始git地址：https://github.com/junegunn/vim-plug
克隆地址： https://gitclone.com/github.com/junegunn/vim-plug
#香港服务器https://doc.fastgit.org ，当前不可用Welcome to leave a message