horseee LLM-Pruner issues

horseee / LLM-Pruner

[NeurIPS 2023] LLM-Pruner: On the Structural Pruning of Large Language Models. Support Llama-3/3.1, Llama-2, LLaMA, BLOOM, Vicuna, Baichuan, TinyLlama, etc.

https://arxiv.org/abs/2305.11627

Apache License 2.0

835 stars 98 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Post training more than 1 epoch leads to performance degradation

#81 sidhantls opened 1 week ago
0
How to prune 20% of parameters?

#80 sidhantls opened 1 week ago
0
evaluate PPL with the post-training model

#79 VincentZ-2020 opened 2 weeks ago
1
关于consecutive_groups

#78 VincentZ-2020 opened 2 weeks ago
0
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8b in position 1

#77 BrownTan closed 3 weeks ago
4
Taylor pruner under-utilizing resources

#76 marianbasti opened 1 month ago
0
Creating custom configuration files in hgging face format

#75 sriyachakravarthy opened 1 month ago
0
No such file or directory: pytorch_model.bin

#74 yaolu-zjut opened 1 month ago
2
Can the pruned model be converted into a gguf file？

#73 pxrgod opened 1 month ago
0
Custom Model pruning

#72 saidineshpola opened 1 month ago
0
Does it support qwen2?

#71 yangxue-1 opened 1 month ago
1
Update ppl_dataset.py with trust_remote_code

#70 manickavela29 opened 1 month ago
1
Llama3 reports shape error after pruning

#69 WentaoTan opened 1 month ago
7
Loading pruned model for causal llm

#68 sriyachakravarthy opened 1 month ago
0
Update hf_prune.py

#67 aritralegndery opened 1 month ago
1
I would like to ask if the current version is suitable for qwen.

#66 wangxiaoxue opened 2 months ago
3
Support GQA &Llama 3 & Llama 3.1.

#65 horseee closed 2 months ago
0
Adaptation of GQA

#64 junzhang-zj opened 2 months ago
7
No pytorch_model.bin file in the tune_log/llama_0.2/checkpoint-200 folder

#63 hebowei2000 opened 3 months ago
3
请问能裁剪普通的transformer模型吗

#62 SKY072410 opened 3 months ago
0
请问可以支持chatglm3剪枝吗

#61 Franklin-L opened 3 months ago
0
Difference in Perplexity Values

#60 nikhil-ghosh-berkeley opened 4 months ago
0
No random seed Settings found in post_training.py

#59 JunKong5 opened 4 months ago
0
Evaluation：UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8b in position 1: invalid start byte

#58 manlenzzz opened 5 months ago
1
Pruning llama3

#57 yinwangsong opened 5 months ago
0
I tired Mistral 7b model, but I got this issue

#56 TejasLidhure opened 5 months ago
0
How to prune the embedding and lm_head?

#55 L-hongbin opened 6 months ago
0
Is this method implementable on multi-GPUs?

#54 LeonCheng0129 opened 6 months ago
0
RecursionError: maximum recursion depth exceeded

#53 Zhenyu001225 opened 6 months ago
2
Unable to reproduce the results for param_first and param_second in the paper after finetuning.

#52 danyal97 opened 7 months ago
0
剪枝率值的问题

#51 qxpBlog closed 8 months ago
0
延迟评估

#50 qxpBlog closed 8 months ago
4
The quantization of the compressed models

#49 lihuang258 opened 8 months ago
0
ConnectionError: Couldn't reach https://raw.githubusercontent.com/wojzaremba/lstm/master/data/ptb.train.txt (ReadTimeout(ReadTimeoutError("HTTPSConnectionPool(host='raw.githubusercontent.com', port=443): Read timed out. (read timeout=100)")))

#48 qxpBlog closed 8 months ago
2
OSError: Can't load tokenizer for 'baffo32/decapoda-research-llama-7B-hf'.

#47 qxpBlog closed 8 months ago
2
Cannot use huggface to load

#46 coderchem opened 9 months ago
0
Issue: Missing Generation of `pytorch_model.bin` File During Model Tuning

#45 WilliamYi96 closed 9 months ago
5
cannot import name 'SiLUActivation' from 'transformers.activations'

#44 azuryl closed 9 months ago
1
401 Client Error: Unauthorized for url: https://huggingface.co/decapoda-research/llama-7b-hf/resolve/main/tokenizer_config.json

#43 azuryl opened 9 months ago
1
Adding a tutorial for adapting new models?

#42 zhichaoxu-shufe opened 10 months ago
0
在将部分层进行剪枝之后，不能直接通过tgi加载模型

#41 coderchem opened 10 months ago
0
Pruning MQA?

#40 jianyuheng opened 10 months ago
0
Question related to the model tuning

#39 shawnricecake opened 10 months ago
2
为什么num_examples默认是10？

#38 coderchem opened 10 months ago
2
the new pytorch.bin is bigger than original model issue

#37 lb553024300 opened 10 months ago
4
Supporting device_map = 'auto' similar to the one in .from_pretrained method from Huggingface

#36 Ahmed-Roushdy opened 10 months ago
3
a post-training issue

#35 cmnfriend opened 10 months ago
2
Reproducing paper results

#34 grigorn opened 10 months ago
6
Latency code

#33 tuidan closed 10 months ago
2
Can not import LlamaConfig

#32 Ahmed-Roushdy closed 11 months ago
1