Skip to content

Issues: deepspeedai/DeepSpeedExamples

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Clear current search query, filters, and sorts
22 Open 16 Closed
22 Open 16 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

During the training of Step 3, the reward score of my language model collapsed to a stable point deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#586 opened Jun 9, 2023 by scarydemon2
7
Evaluation Loader for DeepSpeed Chat Example step 2 (reward model training) deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#581 opened Jun 8, 2023 by harveyp123
step3_rlhf_finetuning and two tokenizers deespeed chat DeepSpeed Chat modeling Related to modeling questions. new-config A modified config from the given example question Further information is requested
#577 opened Jun 6, 2023 by GenVr
@xiaoxiawu-microsoft
3
step2 bug fix for loss = nan when using BLOOM(which is left padding style) deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#571 opened Jun 2, 2023 by scarydemon2
@yaozhewei
messy response from model trained with opt-1.3b and Dahoas/rm-static deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#569 opened Jun 2, 2023 by treya-lin
@awan-10
1
【problem discuss】Critic Loss can not decrease deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#556 opened May 30, 2023 by watermelon-lee
17
what data should I use in step 3 deespeed chat DeepSpeed Chat modeling Related to modeling questions. question Further information is requested
#555 opened May 29, 2023 by scarydemon2
【Need Help】What is [state, action, reward ] in NLP Scenario for PPO in deepspeed-chat deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#552 opened May 26, 2023 by valkryhx
1
step3 answer is not correct deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#547 opened May 25, 2023 by BaiStone2017
5
The min_length setting force the model generate to max length, which produce repeated or nonsense result deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#539 opened May 20, 2023 by TheEighthDay
1
Hyper-param tuning for PPO deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#532 opened May 17, 2023 by luzai
@yaozhewei
1
Rewards in ppo seem to be recomputed many times deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#528 opened May 16, 2023 by dwyzzy
@yaozhewei
2
RLHF model return '{: {: {:' of every input deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#518 opened May 11, 2023 by kuangdao
@xiaoxiawu-microsoft
4
The reward in step3 seems to be completely random without any noticeable increase. deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#489 opened May 7, 2023 by laoda513
@yaozhewei
10
PPO training unable to reproduce the training log provided deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#474 opened May 4, 2023 by REIGN12
@yaozhewei
1
Adding two loss from actor will lead to an error " gradient computed twice for this partition" deespeed chat DeepSpeed Chat modeling Related to modeling questions. new-config A modified config from the given example
#458 opened Apr 28, 2023 by piekey1994
@tjruwase
4
Bug: incorrect metrics evaluating for step two deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#425 opened Apr 25, 2023 by s-isaev
@xiaoxiawu-microsoft
3
step3 failed actor opt_1.3b critic opt_350m Exception: Current loss scale already at minimum - cannot decrease scale anymore. Exiting run deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#419 opened Apr 25, 2023 by BaiStone2017
@xiaoxiawu-microsoft
1
Step2 training get a negative score and accuray is below 60% deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#322 opened Apr 17, 2023 by dlnlpchenliyu
@tjruwase
1
Model performance suprisingly bad deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#318 opened Apr 16, 2023 by ruihan0495
@tjruwase
9
My model Performs Badly...Is GPU memory to small? deespeed chat DeepSpeed Chat modeling Related to modeling questions.
#307 opened Apr 15, 2023 by Trace2333
@tjruwase
12
Does the model in this framework have to be trained to have conversations? deespeed chat DeepSpeed Chat modeling Related to modeling questions. question Further information is requested
#306 opened Apr 15, 2023 by code-isnot-cold
1
ProTip! What’s not been updated in a month: updated:<2025-02-25.