unslothai unsloth issues

unslothai / unsloth

Finetune Llama 3.2, Mistral, Phi, Qwen 2.5 & Gemma LLMs 2-5x faster with 80% less memory

https://unsloth.ai

Apache License 2.0

18.4k stars 1.29k forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Llama 3.2 vision finetuning error (Unsupported: hasattr ConstDictVariable to)

#1325 adi7820 opened 1 hour ago
0
Unsloth Phi-3.5 LoRA: 3x the Number of Trainable Parameters with the Same Hyperparameters

#1324 KristianMoellmann opened 3 hours ago
0
Saving the model with save_pretrained_merged failed.

#1323 WATCHARAPHON6912 opened 7 hours ago
1
Loading a vision lora fails with `ValueError: Unrecognized model in lora_model. Should have a `model_type` key in its config.json`

#1322 saum7800 opened 8 hours ago
3
5 times slower then before

#1321 SidneyLann opened 12 hours ago
4
How to train only last few layers using FastLanguageModel

#1320 gneeraj97 opened 15 hours ago
1
How to fine-tune LLaMA 3.2 11B Vision using LoRA with the recent update?

#1319 yukiarimo opened 19 hours ago
3
Vision

#1318 danielhanchen closed 20 hours ago
0
TypeError: expected string or bytes-like object

#1317 AndersonPedrosa35 closed 19 hours ago
1
Feat/kto

#1316 Erland366 opened 23 hours ago
0
Vision support

#1315 danielhanchen closed 1 day ago
0
failed finetune qwen32b_awq_int4 using lora with llama-factory

#1314 Daya-Jin opened 1 day ago
5
ImportError when importing FastLanguageModel from unsloth

#1313 MurphyJUAN opened 1 day ago
3
The tokenizer does not have a {% if add_generation_prompt %}

#1312 Galaxy-Husky opened 1 day ago
2
Not able to load model from huggingface repo with correct path (FileNotFoundError: invalid repository id)

#1311 ygl1020 opened 1 day ago
1
what was the quantisation algorithm used in unsloth/Llama-3.2-1B-bnb-4bit?

#1310 jayakommuru opened 2 days ago
1
Does tensorRT-LLM support serving 4bit quantised unsloth Llama model

#1309 jayakommuru opened 2 days ago
1
Unable to get the output due to bitsandbytes library error

#1308 iamnikhildogra opened 2 days ago
1
Feat/kaggle-gguf-on-tmp

#1307 Erland366 opened 2 days ago
0
Pythia Models Unsupported

#1306 sert121 opened 3 days ago
1
Is it possible for Unsloth to support naive model parallelism?

#1305 Songjw133 opened 4 days ago
0
About alpha/rank in lora

#1304 Vital1162 opened 4 days ago
3
Dataset for train to translate language

#1303 nichellehouston opened 4 days ago
1
AttributeError: module 'google.protobuf.descriptor' has no attribute '_internal_create_key'

#1302 Sherlock-shy opened 4 days ago
2
Epoch number 3 disappears during evaluation

#1301 daegonYu opened 4 days ago
1
Training Setting

#1300 nichellehouston opened 5 days ago
2
Apple's cross entropy computation

#1299 fzyzcjy closed 5 days ago
2
Can you add support for apple/ml-cross-entropy?

#1298 zfflxx opened 6 days ago
6
`AttributeError: 'LlamaForCausalLM' object has no attribute 'update'`

#1297 scigeek72 opened 6 days ago
2
Error while importing "from unsloth import FastLanguageModel"

#1296 thesillystudent opened 1 week ago
2
Fix too sensitive "Unsloth currently does not support multi GPU setups" when training with a single GPU in a multi-GPU environment.

#1295 giuliabaldini opened 1 week ago
8
Error while importing unsloth in databricks

#1294 BurakaKrishna opened 1 week ago
2
fix/sfttrainer-compatibility

#1293 Erland366 closed 1 week ago
1
Low GPU utilization when running Unsloth-finetuned Qwen2.5-Coder-14B-Instruct-128K-GGUF

#1292 e1ijah1 opened 1 week ago
3
Extremely long context finetuning

#1291 GianlucaDeStefano opened 1 week ago
2
Train on responses only does not seem to work for Mistral format

#1290 LostRuins opened 1 week ago
3
Added Support for Apple Silicon

#1289 shashikanth-a opened 1 week ago
4
Bug fixes

#1288 danielhanchen closed 1 week ago
0
fix indentation error in models/_utils.py:209

#1287 grpathak22 opened 1 week ago
0
Fix orpo/dpo trainer

#1286 dame-cell opened 1 week ago
0
`unexpected keyword argument tokenizer` [FIXED]

#1285 avemio-digital opened 1 week ago
4
`{% if add_generation_prompt %}` [FIXED]

#1284 giuliabaldini opened 1 week ago
6
Error loading model: Unsloth: unsloth/Meta-Llama-3.1-8B-bnb-4bit not supported yet! Make an issue to https://github.com/unslothai/unsloth!

#1283 sree-tejis opened 1 week ago
3
Gradient norm is zero for training Qwen2.5-0.5B-Instruct in unsloth=="2024.11.6"

#1282 joe32140 opened 1 week ago
1
Fix/export mistral

#1281 Erland366 closed 1 week ago
3
Qwen 2.5

#1280 danielhanchen closed 1 week ago
0
FileExistsError: [WinError 183]

#1279 rogersohandsome opened 1 week ago
1
Will the open source version support multiple Gpus later?

#1278 first-li opened 1 week ago
3
Can you support the fine-tuning of the MiniCPM3-4B model ?

#1277 faceair opened 1 week ago
3
fix/sft-trainer

#1276 Erland366 closed 1 week ago
5