tatsu-lab/stanford_alpaca issues and pull requests

#313 - Keyword arguments {'add_special_tokens': False} not recognized.

Issue - State: open - Opened by cswangxiaowei 7 months ago - 1 comment

#155 - [Large Data Training] It can train, but there seems to be a error

Issue - State: closed - Opened by WangRongsheng over 1 year ago - 2 comments

#133 - ValueError: Your setup doesn't support bf16/gpu. You need torch>=1.10, using Ampere GPU with cuda>=11.0

Issue - State: open - Opened by GUORUIWANG over 1 year ago - 9 comments

#100 - Correct prompt.txt mistakes

Pull Request - State: open - Opened by 19245222 over 1 year ago

#99 - A simple codebase for llama finetuning with adapter

Issue - State: open - Opened by gaopengpjlab over 1 year ago - 1 comment

#98 - Update README.md: Grammar

Pull Request - State: open - Opened by yunginnanet over 1 year ago

#97 - Elaborate on used prompt

Issue - State: open - Opened by stefnnn over 1 year ago

#96 - Train using 13b llama model

Issue - State: open - Opened by dev2021-ctrl over 1 year ago - 4 comments

#95 - How to make it work on Google Cloud TPU?

Issue - State: open - Opened by aicheung over 1 year ago - 2 comments

#94 - Exception: Could not find the transformer layer class to wrap in the model.

Issue - State: open - Opened by Cloopen-ReLiNK over 1 year ago - 3 comments

#93 - High training loss of LLaMA 13B

Issue - State: closed - Opened by zwhe99 over 1 year ago - 9 comments

#92 - training on v100

Issue - State: open - Opened by shaileshj2803 over 1 year ago - 10 comments

#91 - Have you ever test the continuous conversation capability of Alpaca?

Issue - State: open - Opened by zhaoshitian over 1 year ago - 1 comment

#90 - Dose anyone have prediction code?

Issue - State: closed - Opened by Hins over 1 year ago - 2 comments

#89 - Andres/refactored translate

Pull Request - State: closed - Opened by andyherfer over 1 year ago

#88 - BBH stats?

Issue - State: open - Opened by i-am-neo over 1 year ago

#87 - Proposal: should we have a slack channel or discord room for issue discussions

Issue - State: open - Opened by bingjie3216 over 1 year ago - 1 comment

#86 - LLaMA-13B (HF) Fails with OOM on a dual A100-80GB

Issue - State: open - Opened by jtang613 over 1 year ago - 1 comment

#85 - Separate training code and dependencies to make who want to fine-tune only easier

Pull Request - State: open - Opened by HUGHNew over 1 year ago

#84 - Initial commit

Pull Request - State: closed - Opened by Ccdzjf over 1 year ago

#83 - Different format for inference ？

Issue - State: open - Opened by PansaLegrand over 1 year ago - 2 comments

#82 - Weights released + frontend, you can try Alpaca 7B here

Issue - State: open - Opened by sergevar over 1 year ago - 1 comment

#81 - A brief summary of the potential issues during the replication and corresponding solutons

Issue - State: open - Opened by puyuanliu over 1 year ago - 2 comments

#80 - No such file or directory: 'LlamaDecoderLayer'

Issue - State: closed - Opened by zachNA2 over 1 year ago - 2 comments

#79 - update notes for training slowdown

Pull Request - State: open - Opened by helloeve over 1 year ago

#78 - Why does the blog mention the PR https://github.com/huggingface/transformers/pull/21955 when it says its merged

Issue - State: open - Opened by nithinhrao over 1 year ago - 2 comments

#77 - Minor: spelling fixes in prompt

Pull Request - State: open - Opened by RikVN over 1 year ago

#76 - OOM when running fine-tune in 4 A100

Issue - State: closed - Opened by Hins over 1 year ago - 2 comments

#75 - Questions about installing Transformer, and the versions of each environment

Issue - State: closed - Opened by 447428054 over 1 year ago - 3 comments

#74 - Content moderation consistently flagging request to count to 100 as inappropriate?

Issue - State: closed - Opened by Patronics over 1 year ago - 1 comment

#73 - finetuning on 3090, is it possible?

Issue - State: open - Opened by yfliao over 1 year ago - 2 comments

#72 - running the project.

Issue - State: closed - Opened by valiantlynx over 1 year ago - 6 comments

#71 - Any plans for using GPT-4 for self-instruct? Or using larger llama models?

Issue - State: open - Opened by JBX2060 over 1 year ago - 2 comments

#70 - Strange inference output

Issue - State: closed - Opened by puyuanOT over 1 year ago - 5 comments

#69 - calculate max_tokens based on prompt tokens

Pull Request - State: open - Opened by bartman081523 over 1 year ago

#68 - Update stanford_alpaca to use transformers main branch

Pull Request - State: closed - Opened by Danivilanova over 1 year ago - 1 comment

#67 - Help with CUDA error: invalid device ordinal

Issue - State: closed - Opened by GooDRomka over 1 year ago - 3 comments

#66 - What are the requirements for the model

Issue - State: closed - Opened by moh21amed over 1 year ago - 3 comments

#65 - OOM after the last epoch

Issue - State: closed - Opened by puyuanliu over 1 year ago - 6 comments

#64 - Solve BUG:AttributeError: module transformers has no attribute LLaMATokenizer

Issue - State: open - Opened by XuyaoWang over 1 year ago - 12 comments

#63 - gpt-3.5-turbo?

Issue - State: open - Opened by jordancole21 over 1 year ago

#62 - Will you release data collected on demo page ?

Issue - State: closed - Opened by diimdeep over 1 year ago

#61 - Resuming from checkpoint

Issue - State: open - Opened by KurtFeynmanGodel over 1 year ago - 9 comments

#60 - Plan to release the web demo code

Issue - State: closed - Opened by testplop over 1 year ago - 1 comment

#59 - Resuming from checkpoint

Issue - State: closed - Opened by KurtFeynmanGodel over 1 year ago - 2 comments

#58 - Exception: Could not find the transformer layer class to wrap in the model

Issue - State: closed - Opened by Cloopen-ReLiNK over 1 year ago - 11 comments

#57 - Due to OOM, who can finetune LLaMA using bitsandbytes for an 8-bit setting on a single 3090?

Issue - State: closed - Opened by linhduongtuan over 1 year ago

#56 - CUDA out of memory for a single core A100 80G GPU

Issue - State: open - Opened by leondelee over 1 year ago - 11 comments

#55 - Update requirements.txt

Pull Request - State: closed - Opened by adarsh057 over 1 year ago - 1 comment

#54 - Numpie lost Factories

Issue - State: closed - Opened by adarsh057 over 1 year ago - 1 comment

#53 - Confusion about input ids

Issue - State: closed - Opened by fuxuliu over 1 year ago - 1 comment

#52 - how to fine-tune on V100

Issue - State: closed - Opened by Morxrc over 1 year ago - 3 comments

#51 - Generation problem after / before instruction fine-tuning

Issue - State: closed - Opened by hxssgaa over 1 year ago - 11 comments

#50 - No evaluation dataset was given for the trainer

Issue - State: closed - Opened by XuhuiRen over 1 year ago - 5 comments

#49 - does support multi-turn training data?

Issue - State: closed - Opened by trouble-maker007 over 1 year ago - 3 comments

#48 - How to inference after finetuning ?

Issue - State: closed - Opened by kriskrisliu over 1 year ago - 20 comments

#47 - Reduce the length of your prompt.

Issue - State: open - Opened by 19245222 over 1 year ago - 1 comment

#46 - OOM issue

Issue - State: closed - Opened by puyuanliu over 1 year ago - 14 comments

#45 - Do you shift the output label?

Issue - State: closed - Opened by gaopengpjlab over 1 year ago - 3 comments

#44 - OpenAIError Error communicating with OpenAI

Issue - State: closed - Opened by 19245222 over 1 year ago - 2 comments

#43 - Loading llama-7b from huggingface

Issue - State: closed - Opened by puyuanliu over 1 year ago - 4 comments

#42 - No checkpoint and no eval_dataset

Issue - State: closed - Opened by kriskrisliu over 1 year ago - 1 comment

#41 - Comparing training log [Shared my training log]

Issue - State: open - Opened by charliezjw over 1 year ago - 3 comments

#40 - Can you share the log of your finetuning code?

Issue - State: open - Opened by gaopengpjlab over 1 year ago - 1 comment

#39 - CUDA out of memory

Issue - State: closed - Opened by waterhorse1 over 1 year ago - 9 comments

#38 - Bigger LLaMA models

Issue - State: closed - Opened by alexl83 over 1 year ago - 1 comment

#37 - why 52K?

Issue - State: closed - Opened by i-am-neo over 1 year ago - 1 comment

#36 - How to train with the Bible content?

Issue - State: open - Opened by paulocoutinhox over 1 year ago - 3 comments

#35 - inference kwargs

Issue - State: closed - Opened by 1024er over 1 year ago - 4 comments

#34 - Update the README instructions, especially the PR install command included

Pull Request - State: closed - Opened by pervrosen over 1 year ago - 3 comments

#33 - Question about training precision

Issue - State: closed - Opened by 152334H over 1 year ago - 1 comment

#32 - Fine-Tuning very slow (6h->24h??)

Issue - State: closed - Opened by chavinlo over 1 year ago - 49 comments

#31 - Not quite understand the importance of this repo.

Issue - State: closed - Opened by 19245222 over 1 year ago - 1 comment

#30 - [DEV] upload sanitized training code

Pull Request - State: closed - Opened by lxuechen over 1 year ago

#29 - Any APIs like OpenAI will be released in the future?

Issue - State: closed - Opened by samchen8008 over 1 year ago - 1 comment

#28 - When can we support airgap installation?

Issue - State: closed - Opened by samchen8008 over 1 year ago - 1 comment

#27 - could be open source the model ?

Issue - State: closed - Opened by ucas010 over 1 year ago - 4 comments

#26 - Inquiry: Inference Parameters used for Gradio Demo

Issue - State: closed - Opened by TheZennou over 1 year ago - 2 comments

#25 - Public release of model weights

Issue - State: closed - Opened by topiconcept over 1 year ago - 3 comments

#24 - minor: fix numbering in prompt

Pull Request - State: closed - Opened by dooart over 1 year ago - 3 comments

#23 - Finetuning using standard hugging face training code

Issue - State: closed - Opened by urstrulyajay over 1 year ago - 2 comments

#22 - [Python3.8]fix: TypeError: 'type' object is not subscriptable

Pull Request - State: closed - Opened by fuhengwu2021 over 1 year ago - 1 comment

#21 - [Q] How much vRAM does finetuning LLaMa 7B require?

Issue - State: closed - Opened by NightMachinery over 1 year ago - 3 comments

#20 - Add output to finetuning prompt

Pull Request - State: closed - Opened by lewtun over 1 year ago - 1 comment

#19 - generate_instruction_following_data

Issue - State: closed - Opened by chlee29 over 1 year ago - 1 comment

#18 - Update README.md

Pull Request - State: closed - Opened by eltociear over 1 year ago

#17 - Update alpaca_data.json math issues

Pull Request - State: closed - Opened by NPap0 over 1 year ago - 1 comment

#16 - Are some layers frozen while fine-tuning？

Issue - State: closed - Opened by TccccD over 1 year ago - 3 comments

#15 - Update prompt.txt

Pull Request - State: closed - Opened by Suro-One over 1 year ago - 1 comment

#14 - Training recipe??

Issue - State: closed - Opened by milsun over 1 year ago - 4 comments

#13 - Can this one support MacOS? Any particular hardware is required?

Issue - State: closed - Opened by samchen8008 over 1 year ago - 1 comment

#12 - Fix alpaca_data.json math

Pull Request - State: closed - Opened by bstst over 1 year ago - 1 comment

GitHub / tatsu-lab/stanford_alpaca issues and pull requests