LLMZoo
LLMZoo copied to clipboard
FreedomIntelligence
⚡LLM Zoo is a project that provides data, models, and evaluation benchmark for large language models.⚡
The files in the checkpoints are small, 266M in total. Cannot directly loading from it.
消费级显卡：3060 - 12GB 启动命令：python -m llmzoo.deploy.cli --model-path FreedomIntelligence/phoenix-inst-chat-7b --max-gpu-memory 10Gib --load-8bit 在 Human 输入 prompt 回车后内存不足： ``` Human: Who are you? Assistant: ``` OutOfMemoryError: CUDA out of memory. Tried to...
Hello, I run the script train_phoenix_7b.sh to save the checkpoint. When I run the script for the second time to resume from the checkpoint, an error is reported. I don’t...
Nice work on your repo. I'm curious to see whether you managed to achieve your results in Chinese and multilingual models with or without resizing the vocabulary size of the...
我用一张A100-SXM4-40GB尝试finetue时，将batchsize设置成1，但内存还是爆了。   难道单卡A100真的跑不起来吗
When I used the "pip install -r requirements.txt", then I got the following probem: " File "/tmp/pip-build-env-4dcxq614/overlay/lib/python3.9/site-packages/torch/utils/cpp_extension.py", line 1909, in _run_ninja_build raise RuntimeError(message) from e RuntimeError: Error compiling objects for...
On the README page you say that LLaMA and Alpaca are open source. Since they are restricted for non commercial, research only they can not be called open-souce.
how to fix this error：Tokenizer class BloomTokenizer does not exist or is not currently imported.  my transformers version is 4.30.0.dev0.
你好，在模型评估的代码里遇到以下问题： 1、执行流程是否为 run_evaluation_gpt35.sh -> run_stat_evaluetion.sh，并保证”eva(score/oder)“一致 2、缺少 roleplay-questions-zh.jsonl 文件 3、如何评估自己的模型 还望百忙之中予以答复，不胜感激
Metadata
Owner
Metadata
⚡LLM Zoo is a project that provides data, models, and evaluation benchmark for large language models.⚡