issues
search
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
18.37k
stars
1.28k
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Resizing tokenizer leads to missing end token and garbage response?
#1273
Mark-DelGrande
opened
1 week ago
1
Jupyter notebook: No module named 'unsloth'
#1272
iwouldratherbeatthebeach
opened
1 week ago
3
dataset for train model to translate language
#1271
nichellehouston
closed
1 week ago
1
feat: add option for using ADOPT optimizer based on Taniguchi, Shohei, et al.
#1270
Selich
opened
1 week ago
1
DOC Update - Update README.md with os.environ in example
#1269
udaygirish
closed
1 week ago
1
cannot load some models via vllm
#1268
yananchen1989
opened
1 week ago
11
save_pretrained_merged ruins my model
#1267
Romiroz
opened
1 week ago
3
Couldn't build proto file into descriptor pool! Invalid proto descriptor for file "sentencepiece_model.proto": sentencepiece_model.proto: A file with this name is already in the pool.
#1266
CurtiusSimplus
opened
2 weeks ago
9
LoRA on Qwen 2.5 does not patch qkv matrices
#1265
MinghaoYan
opened
2 weeks ago
1
TypeError: SFTTrainer.__init__() got an unexpected keyword argument 'dataset_text_field'
#1264
officialsahyaboutorabi
closed
1 week ago
6
`triton.language' has no attribute cast` [FIXED]
#1263
arianyambao
opened
2 weeks ago
11
`train_on_responses_only` doesn't work for Mistral models
#1262
XiaomoWu
opened
2 weeks ago
5
CAN'T LOAD: AttributeError: 'LlamaForCausalLM' object has no attribute 'update'
#1261
yukiarimo
opened
2 weeks ago
10
[FIXED] `dtype c10::BFloat16, but got float`
#1260
CurtiusSimplus
closed
2 weeks ago
8
Bug fixes
#1259
danielhanchen
closed
2 weeks ago
0
ValueError: Unsloth: Untrained tokens of [[128004]] found
#1258
Hyfred
opened
2 weeks ago
2
Getting "compiled_autograd.enable() requires no threads in backwards()" on running SFTTrainer on unsloth/gemma-2 models
#1257
sudha-kannan
closed
2 weeks ago
0
fix/autograd_compile
#1256
Erland366
closed
2 weeks ago
2
Bug fixes
#1255
danielhanchen
closed
2 weeks ago
0
Fix: cast logits to float32 in cross_entropy_forward to prevent errors
#1254
Erland366
closed
2 weeks ago
2
Problem with installing packages and dependencies (triton in particular)
#1253
AllenY687
closed
2 weeks ago
1
import unsloth causes error: pip install unsloth-zoo
#1252
DaddyCodesAlot
opened
2 weeks ago
5
issue when using default settings for training
#1251
Ammar-Alnagar
closed
1 week ago
5
[FIXED] `compiled_autograd.enable()` Gemma
#1250
InderjeetVishnoi
closed
2 weeks ago
4
Bug fix
#1249
danielhanchen
closed
2 weeks ago
1
`AssertionError('initial value for logits` error [FIXED]
#1248
daegonYu
opened
2 weeks ago
10
Errors occurring in Pip Installation : torch 2.5 and CUDA 12.4
#1247
daegonYu
closed
2 weeks ago
1
fix/get_chat_template
#1246
Erland366
closed
1 week ago
1
Bug fixes
#1245
danielhanchen
closed
2 weeks ago
0
RuntimeError: CUDA error during inference from saved lora weights
#1244
danisharoonds
opened
2 weeks ago
1
Dataset creation to use with unsloth fine tuning
#1243
gaussiangit
closed
6 days ago
1
Unsloth error unable to push to hub
#1242
hung-ngm
closed
1 week ago
2
how to only do lora on the lm_head?
#1241
brando90
opened
2 weeks ago
3
why is unsloth thinking I'm doing multi gpu optimization when I'm not?
#1240
brando90
opened
2 weeks ago
3
Fine tuned Llama3.1 does not support tools
#1239
darkroasted
opened
2 weeks ago
1
erorr
#1238
werruww
opened
2 weeks ago
5
RuntimeError: `ptxas` failed with error code 4294967295:
#1237
heiheiheibj
opened
2 weeks ago
2
Throw error when inferencing longer than max_popsition_embeddings
#1236
Datta0
closed
2 weeks ago
0
CLI now handles user input strings for dtype correctly
#1235
Rabbidon
closed
2 weeks ago
1
Which Torch & Python
#1234
IzzyHibbert
closed
2 weeks ago
5
Overlap matrix multiplication (needs tensor core) and other things like activation (needs cuda core and memory bandwidth) to speed up
#1233
fzyzcjy
opened
2 weeks ago
2
AttributeError: 'torchvision' has no attribute 'extension' When Using Unsloth on Kaggle
#1232
Saber120
closed
2 weeks ago
1
Unsloth error with trl 0.11.4
#1231
mohit-raghavendra
closed
2 weeks ago
7
Why is memory bandwidth only half used? Is it possible we speed up by utilizing this?
#1230
fzyzcjy
opened
2 weeks ago
2
Is it possible to use `train_on_responses_only` with the Mistral template?
#1229
kldzj
opened
2 weeks ago
2
support
#1228
Qarqor5555555
opened
2 weeks ago
3
Remove "embed_tokens" and "lm_head" Lora layers when loading CPT trained models
#1227
daegonYu
closed
2 weeks ago
2
Update README.md
#1226
WontonSam
closed
3 weeks ago
1
fix/load-checkpoint-add-new-tokens
#1225
Erland366
opened
3 weeks ago
3
OSError: could not get source code when loading a model using a for loop
#1224
daegonYu
opened
3 weeks ago
4
Previous
Next