xfactlab orpo issues - Githubissues

xfactlab / orpo

Official repository for ORPO

Apache License 2.0

412 stars 38 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Ignore

#36 xlar-sanjeet closed 1 month ago
0
compute_logps function, why does it return also prob for the last token of answer

#35 javismiles closed 2 months ago
7
Could you please tell me which OpenAI API you used during the MT-Bench evaluation?

#34 hitszxs opened 2 months ago
2
[Question] how does orpo combine dpo into sft ?

#33 wj-Mcat opened 3 months ago
1
Discarding the prompt tokens only with the positive labels and not with the negative ones

#32 javismiles closed 3 months ago
2
Recreating the setup with CUDA 12.1; grad norm is nan

#31 Jayant1234 closed 2 months ago
2
[Question] ORPO Fine-tuning Data Format

#30 nooobodynose closed 3 months ago
2
Add-log1p

#29 jiwooya1000 closed 4 months ago
1
Unexpected results using ORPO trl

#28 celsowm closed 4 months ago
4
Poor performance on llama3

#27 JasonZhu1313 opened 5 months ago
3
Can you share your training loss log for Mistral-ORPO-β (7B)?

#26 yechenzhi closed 2 months ago
11
Is there any statement error in 4.2 of the paper？

#25 Chic-star closed 5 months ago
2
Update main.py

#24 ganeshkrishnan1 opened 5 months ago
0
no reference model?

#23 kxleee closed 4 months ago
1
Memory Consumption

#22 paulcx closed 5 months ago
1
what's the difference between alpha and beta version?

#21 Labmem009 closed 5 months ago
1
Can you please give us a guide line to use your method to train on the LLAMA factory?

#20 TonyQJH closed 2 months ago
1
Added quantization options

#19 javafa opened 5 months ago
3
Loss device for ORPOTrainer

#18 ganeshkrishnan1 opened 5 months ago
16
attention mask in compute_logps function

#17 hjc3613 closed 5 months ago
3
Add disable_prompt_loss argument

#16 nlee-208 closed 5 months ago
0
Fix `test_orpo_trainer_demo.py`

#15 alvarobartt closed 5 months ago
1
Add `seed` in `default_args` for reproducibility

#14 alvarobartt closed 5 months ago
1
[Question] Memory requiremsnts for ORPOTrainer

#13 snassimr closed 5 months ago
1
how to install requirements.txt on colab?

#12 srn-source closed 5 months ago
3
how to do ORPO with ShareGPT data?

#11 pabl-o-ce closed 5 months ago
2
[Question] ORPO + SFTTrainer + QLora

#10 snassimr closed 5 months ago
1
Run with Deepspeed raise ERROR: Gradient computed twice for this partition

#9 iFe1er closed 5 months ago
3
Add llama-factory to readme

#8 hiyouga closed 6 months ago
1
Doubt about the formatting of the `prompt`, `chosen` and `rejected`

#7 alvarobartt closed 6 months ago
2
Suggestion: Calculate NLL Loss on neutral preference set.

#6 linux-leo closed 6 months ago
2
prompt formatting issue

#5 RonanKMcGovern closed 6 months ago
12
Questions regarding the reproduction of Mistral's results on mt-bench

#4 JasonZhu1313 closed 6 months ago
11
4xA6000 training - failed to save model

#3 rkinas closed 6 months ago
7
Add new .sh and support for general pair datasets

#2 jiwooya1000 closed 6 months ago
0
Support other preference datasets

#1 jiwooya1000 closed 6 months ago
0