MiniCPM
MiniCPM copied to clipboard
MiniCPM4: Ultra-Efficient LLMs on End Devices, achieving 5+ speedup on typical end-side chips
minor fix README.md fro llama.cpp
### Description / 描述 from transformers import AutoModelForCausalLM, AutoTokenizer import torch import time s = time.time() path = "D:/MiniCPM3-4B" device = "cpu" tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained(path, torch_dtype=torch.bfloat16,...
### Feature request / 功能建议 可否提供一个 无限长上下文 MapReduce 的 sample code?
### Description / 描述 cos = cos[position_ids].unsqueeze(unsqueeze_dim) # [bs, 1, seq_len, dim] IndexError: index is out of bounds for dimension with size 0 ### Case Explaination / 案例解释 cos =...
### Feature request / 功能建议 可否提供正確的CUDA version, PyTorch version, 以及 deepspeed version?
Congrats to your great work!
### Description / 描述 使用vllm部署: python -m vllm.entrypoints.openai.api_server --model /data2/MiniCPM --host 0.0.0.0 --port 10999 --max-model-len 2048 --served-model-name minicpm --trust_remote_code 用evol-scope进行并发测试: 结果很慢,100并发结果: Benchmarking summary: Time taken for tests: 384.675 seconds Expected...
### Description / 描述 非bug,请教一下你们预训练用了1.1T tokens,花了多少GPU和时间@LDLINGLINGLING ### Case Explaination / 案例解释 _No response_