Skip to content

Issues: huggingface/open-r1

How to contribute
#23 opened Jan 25, 2025 by lewtun
Open 9
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

164 Open 55 Closed
164 Open 55 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

How many resources are required to train deepseek r1 671b using grpo?
#413 opened Feb 24, 2025 by LiuShixing
Why does the reward keep rising while the length keeps fluctuating horizontally?
#412 opened Feb 24, 2025 by zhufz
Stuck at lighteval "COMPUTING METRICS"
#411 opened Feb 24, 2025 by chenjingming99
GRPO trainer error
#410 opened Feb 24, 2025 by xiangxinhello
Why is the result variable when using vllm for inference, even though I’ve set the temperature to 0.01? The result should be fixed, right?
#409 opened Feb 24, 2025 by hellen9527
Evaluation : TypeError: VLLMModelConfig.__init__() got an unexpected keyword argument 'temperature:0.6'
#408 opened Feb 24, 2025 by ChenDRAG
2
GRPO reward hacking
#407 opened Feb 24, 2025 by Kidand
1
train/loss and rewards/format_reward are always '0.0' while running the grpo step
#405 opened Feb 23, 2025 by yfliao
When I compute loss, I met logits=None, past_key_values=None, hidden_states=None, attentions=None
#404 opened Feb 23, 2025 by royal-dargon
1
Instead of rising steadily, the reward fluctuates wildly
#403 opened Feb 23, 2025 by Hasuer
6
Error with boardcast completion ids
#400 opened Feb 23, 2025 by hhnqqq
1
[SFT] Model trainned on Bespoke-Stratos-17k get really low performence on Math-500
#396 opened Feb 22, 2025 by ListentoMe0112
7
less strict format reward function is necessary for early training and the default system prompt is not appropriate for answers like '<think>...<answer>'
#395 opened Feb 22, 2025 by jssuncx
After I trained for 500 steps, the length of think became smaller and smaller, and even disappeared.
#394 opened Feb 22, 2025 by yaxundai
Errors in lighteval math500/gpqa evaluation about Sympify parse problem
#391 opened Feb 22, 2025 by Larry919
About the sampler problems.
#390 opened Feb 21, 2025 by linkedlist771
3
sft with multiple GPUs got stuck
#389 opened Feb 21, 2025 by SharonJin422
2
Issue with training 7B model
#383 opened Feb 20, 2025 by rqzhangberkeley
evaluation error when using origin qwen due to max_model_length
#382 opened Feb 20, 2025 by glennccc
1
how to set sampling parameters when do evaluation
#381 opened Feb 20, 2025 by ItGirls
2
How to set cuda device for your Data generation pipline
#380 opened Feb 20, 2025 by Aristo23333
Cannot find tasks extended|lcb:codegeneration in task list or in custom task registry
#378 opened Feb 20, 2025 by wccccp
2
Error Encountered During livecodebench eval Execution
#375 opened Feb 19, 2025 by JackLingjie
The kl divergence collapses but the format reward becomes larger
#373 opened Feb 19, 2025 by yuki-younai
1
DeepSpeed ZeRO-3 Causes Instance Crash on Large max_completion_length in GRPO Training
#372 opened Feb 19, 2025 by troy12x
1
ProTip! Mix and match filters to narrow down what you’re looking for.