Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / allenai/RL4LMs issues and pull requests
#73 - Which version of pip should we use when we try to set up the environment?
Issue -
State: open - Opened by Wwwendyy 13 days ago
- 1 comment
#72 - An issue happened when I set up
Issue -
State: open - Opened by Wwwendyy 15 days ago
#71 - Question about the classifier used for IntentAccuracyDailyDialog.
Issue -
State: open - Opened by zhangjf-nlp 5 months ago
#70 - Migrate to current version of gymnasium, SB3, and other libraries.
Pull Request -
State: open - Opened by Kripner 7 months ago
#69 - Upgrade to torch 2.0
Issue -
State: open - Opened by agastyaseth 8 months ago
- 1 comment
#68 - how to stop env parallel multi-process to debug env.step()?
Issue -
State: open - Opened by invoker-LL 9 months ago
#67 - Trying to use rl4lm with more recent libraries
Pull Request -
State: closed - Opened by JosefSlavicek 10 months ago
#66 - Is PPO really better than SFT (in general)? under the condition of same amount of data
Issue -
State: open - Opened by allanj 10 months ago
- 1 comment
#65 - Do you have any plans to apply the recently published Reinforced Self-Training (ReST)?
Issue -
State: open - Opened by missflash about 1 year ago
#64 - Pip install error with gym and torch
Issue -
State: open - Opened by BaleChen about 1 year ago
- 3 comments
#63 - NLPO Code Error and Query About gymnasium vs gym Usage
Issue -
State: open - Opened by jinyilun718 about 1 year ago
#62 - Reproducing existing results on NarrativeQA
Issue -
State: open - Opened by yxk23 about 1 year ago
#61 - Memory issue in metric evals?
Issue -
State: open - Opened by AnujMahajanOxf about 1 year ago
#60 - is multi-dimensional reward supported?
Issue -
State: open - Opened by zabir-nabil over 1 year ago
#59 - CPU Support Minor Bug
Issue -
State: open - Opened by tedmoskovitz over 1 year ago
#58 - Fix IndexError when loading checkpoints
Pull Request -
State: open - Opened by Runingtime over 1 year ago
#57 - model.generate.scores returning two scores
Issue -
State: open - Opened by debjitpaul over 1 year ago
#56 - 'GPT2Model' object has no attribute 'first_device'
Issue -
State: open - Opened by Stephanehk over 1 year ago
#55 - Using GPT-2
Issue -
State: open - Opened by oroojlooy over 1 year ago
#54 - How can I inference data with the model after PPO training?
Issue -
State: open - Opened by RyanYip-Kat over 1 year ago
#53 - Bug while loading t5 base model
Issue -
State: open - Opened by Sahajtomar over 1 year ago
- 1 comment
#52 - Error with Accelerate integration + NLPO
Issue -
State: open - Opened by avacaondata over 1 year ago
- 1 comment
#51 - [Question] End-to-end example
Issue -
State: open - Opened by farrokhsiar over 1 year ago
#50 - Fix nlpo configs
Pull Request -
State: closed - Opened by rajcscw over 1 year ago
#49 - In the paper, what is the detail setting of supervised learning? Is SL has additional supervised data?
Issue -
State: open - Opened by guotong1988 over 1 year ago
#48 - Resuming from checkpoint is potentially problematic for IMDB since the splits are resampled
Issue -
State: closed - Opened by zhixuan-lin over 1 year ago
- 1 comment
#47 - `train` and `val` splits are not disjoint for IMDB
Issue -
State: closed - Opened by zhixuan-lin over 1 year ago
- 3 comments
#46 - A question bother me a long time: What is the difference between RL-for-text-generation and delete-0-reward-model-predictions?
Issue -
State: open - Opened by guotong1988 over 1 year ago
#45 - A question bother me a long time: What is the difference between RL-for-text-generation and delete-0-reward-model-predictions?
Issue -
State: closed - Opened by guotong1988 over 1 year ago
- 1 comment
#44 - Bloom Supporting
Issue -
State: open - Opened by c-box over 1 year ago
- 3 comments
#43 - Error when trying to load a checkpoint from Transformers after RL training
Issue -
State: closed - Opened by avacaondata over 1 year ago
- 5 comments
#42 - Metric version incompatible
Issue -
State: open - Opened by c-box over 1 year ago
#41 - _pickle.UnpicklingError: pickle data was truncated
Issue -
State: open - Opened by Oxtay over 1 year ago
#40 - Pip install fix
Pull Request -
State: open - Opened by kolbytn over 1 year ago
#39 - Make sure transformer return past_key_values
Pull Request -
State: open - Opened by DvHuang over 1 year ago
#38 - Value is not broadcastable with batch_shape+event_shape
Issue -
State: open - Opened by vcvcvnvcvcvn over 1 year ago
#37 - Persistent Variance in IMDB
Issue -
State: open - Opened by mnoukhov over 1 year ago
- 1 comment
#36 - fix: OnPolicyAlgorithm doesnot have the parameter: create_eval_env
Pull Request -
State: open - Opened by hscspring over 1 year ago
- 1 comment
#35 - Gradient Accumulation feature proposal
Pull Request -
State: closed - Opened by eublefar over 1 year ago
#34 - Problem with BLEURT reward function
Issue -
State: open - Opened by eublefar over 1 year ago
#33 - Is it possible to release the code based on Jax
Issue -
State: open - Opened by sglucas over 1 year ago
#32 - Evaluating a specific checkpoint
Issue -
State: open - Opened by lovodkin93 over 1 year ago
- 5 comments
#31 - UnderStand Mask model to _get_action_masks in LogitsProcessor
Issue -
State: closed - Opened by xesdiny over 1 year ago
#30 - 'BartForConditionalGeneration' has no attribute 'encoder'
Issue -
State: closed - Opened by keeganstoner over 1 year ago
- 4 comments
#29 - Mix-Precision training
Issue -
State: open - Opened by lovodkin93 over 1 year ago
- 2 comments
#28 - Reproducing IMDB results
Issue -
State: open - Opened by mnoukhov over 1 year ago
- 4 comments
#27 - Is the construction of _value_model necessary?
Issue -
State: closed - Opened by xesdiny over 1 year ago
- 2 comments
#26 - passing extra variable to the forward function
Issue -
State: open - Opened by lovodkin93 over 1 year ago
- 1 comment
#25 - Problems with models that don't have the parallelize() function
Issue -
State: open - Opened by lovodkin93 over 1 year ago
- 1 comment
#24 - Changed from logging with the root logger
Pull Request -
State: open - Opened by JulesGM almost 2 years ago
#23 - Off-policy RL algorithms support
Issue -
State: open - Opened by Div99 almost 2 years ago
- 5 comments
Labels: enhancement, help wanted
#22 - Just a warning that the package doesn't work with Transformers 4.25.1
Issue -
State: open - Opened by JulesGM almost 2 years ago
- 1 comment
#21 - Larger models like GPT-J and GPT-NeoX-20B
Issue -
State: open - Opened by loganlebanoff almost 2 years ago
- 3 comments
#20 - make logo work better on dark theme
Pull Request -
State: closed - Opened by jmhessel almost 2 years ago
#19 - templates and example inputs for mechanical turk
Pull Request -
State: closed - Opened by jmhessel almost 2 years ago
#18 - Implementing self-play
Issue -
State: open - Opened by eublefar almost 2 years ago
- 3 comments
#17 - Added ability to set the log level in a backwards compatible way
Pull Request -
State: closed - Opened by JulesGM almost 2 years ago
- 2 comments
#16 - 100% likely that two function parameters have been merged by accident
Issue -
State: open - Opened by JulesGM almost 2 years ago
- 1 comment
Labels: good first issue, code enhancement
#15 - correcting double tokenizing
Pull Request -
State: closed - Opened by JulesGM almost 2 years ago
- 1 comment
#14 - Fix seq2seq
Pull Request -
State: closed - Opened by rajcscw almost 2 years ago
#13 - Self-designed model
Issue -
State: closed - Opened by AIUSRTMP almost 2 years ago
- 2 comments
#12 - OOM on summarization example
Issue -
State: open - Opened by gabrielhuang almost 2 years ago
- 15 comments
#11 - large difference between val and test on CommonGEN
Issue -
State: closed - Opened by wenting-zhao almost 2 years ago
- 2 comments
#10 - BART supervised
Issue -
State: open - Opened by talent404 almost 2 years ago
- 2 comments
#9 - Error encountered in running the scripts at the read me
Issue -
State: closed - Opened by promiseve almost 2 years ago
- 4 comments
#8 - Some questions about n_steps,n_envs and padding_side.
Issue -
State: closed - Opened by drxmy almost 2 years ago
- 3 comments
#7 - Top-K and Top-p sampling
Issue -
State: open - Opened by boblee22 almost 2 years ago
- 1 comment
#6 - fix bug for nlpo policy for mt5 model
Pull Request -
State: closed - Opened by tatiana-iazykova almost 2 years ago
- 1 comment
#5 - Numbeams
Issue -
State: open - Opened by tatiana-iazykova almost 2 years ago
- 5 comments
Labels: bug, beam_search
#4 - Any plans for Deepspeed/Accelerate integration?
Issue -
State: open - Opened by Breakend almost 2 years ago
- 9 comments
#3 - CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`
Issue -
State: closed - Opened by tatiana-iazykova almost 2 years ago
- 4 comments
#2 - Correct the path to t5_ppo.yml in Readme
Pull Request -
State: closed - Opened by akifumi-wachi-4 almost 2 years ago
- 1 comment
#1 - Readme update
Pull Request -
State: closed - Opened by jmhessel almost 2 years ago