Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / kzl/decision-transformer issues and pull requests
#78 - Number of training steps
Issue -
State: open - Opened by mhngu23 about 1 month ago
#77 - RuntimeError: CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when calling `cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc)`
Issue -
State: open - Opened by LLL-YUE 3 months ago
#76 - Why does the training loss only consider actions and not rewards or state
Issue -
State: closed - Opened by LMCAV 3 months ago
- 1 comment
#75 - Trey - evaluate episode rtg
Pull Request -
State: closed - Opened by TTF2050 5 months ago
#74 - Where are the datasets?
Issue -
State: closed - Opened by wilhem 5 months ago
- 1 comment
#73 - Data to trajectory
Pull Request -
State: closed - Opened by DerrickhSun 6 months ago
#72 - compilation error (MuJoCo)
Issue -
State: open - Opened by NagarajJ111 6 months ago
- 1 comment
#71 - MultiDiscrete Action Space
Issue -
State: closed - Opened by hamzaali98 8 months ago
- 1 comment
#70 - What should return to go be during inferencing?
Issue -
State: closed - Opened by FinAminToastCrunch 8 months ago
- 2 comments
#69 - Fix bugs for return_preds and state_preds
Pull Request -
State: open - Opened by nawta 8 months ago
#68 - Clarified that return_preds were unnecessary
Pull Request -
State: open - Opened by nawta 8 months ago
#67 - Question about the output of the decision transformer
Issue -
State: closed - Opened by Pulsar110 9 months ago
- 3 comments
#66 - Bug fix
Pull Request -
State: closed - Opened by pwyq 11 months ago
#65 - Image Data
Issue -
State: open - Opened by AsadMir10 about 1 year ago
#64 - position embeddings do not vary between various time steps
Issue -
State: closed - Opened by udaymallappa about 1 year ago
- 1 comment
#63 - how to print accurate in this code?
Issue -
State: open - Opened by YuZheng23 about 1 year ago
- 2 comments
#62 - torch.stack bug
Issue -
State: open - Opened by ChaselWan about 1 year ago
#61 - why sample from multinomial distribution during evaluation in Atari?
Issue -
State: open - Opened by sallyqiansun over 1 year ago
#60 - Gsutil Error
Issue -
State: open - Opened by ChaselWan over 1 year ago
- 1 comment
#59 - Lsf dev
Pull Request -
State: open - Opened by Fanshaoliu over 1 year ago
#58 - rerun dt
Pull Request -
State: closed - Opened by Fanshaoliu over 1 year ago
#57 - jskang-230403
Pull Request -
State: closed - Opened by jesnk over 1 year ago
#56 - AttributeError: 'GPT2Config' object has no attribute 'n_ctx'
Issue -
State: closed - Opened by NuvoleY over 1 year ago
- 6 comments
#55 - Questions about dataset preprocessing
Issue -
State: open - Opened by typoverflow over 1 year ago
#54 - [IDEA] Code for dataset generation
Issue -
State: open - Opened by JustinS6626 over 1 year ago
- 1 comment
#53 - Bug in state and action prediction
Issue -
State: open - Opened by ezhang7423 over 1 year ago
#52 - Concept test v3 dualtrain
Pull Request -
State: closed - Opened by nkolln over 1 year ago
#51 - Global position embedding and timesteps look wrong in atari
Issue -
State: closed - Opened by nzw0301 almost 2 years ago
- 2 comments
#50 - Why the padding is different for state, action, reward?
Issue -
State: open - Opened by CeyaoZhang almost 2 years ago
#49 - The setting of final token
Issue -
State: open - Opened by IpadLi almost 2 years ago
#48 - Is there a code for the shortest path search?
Issue -
State: open - Opened by geon0325 almost 2 years ago
- 1 comment
#47 - Any reason for using different models for gym and atari experiments?
Issue -
State: closed - Opened by GilgameshD about 2 years ago
- 2 comments
#46 - Atari results
Issue -
State: closed - Opened by TongZhangTHU about 2 years ago
- 4 comments
#45 - MemoryError: Unable to allocate 6.57 GiB for an array with shape (7056000000,) and data type uint8
Issue -
State: open - Opened by leeruibin about 2 years ago
- 1 comment
#44 - No registered env with id: halfcheetah-medium-v2
Issue -
State: open - Opened by boykac about 2 years ago
- 4 comments
#43 - I am wondering if the shortest path case is included in the code? thanks
Issue -
State: open - Opened by bradley-code-again about 2 years ago
#42 - The results on Mujoco reported in paper might be heavily influenced by env version
Issue -
State: open - Opened by linprophet over 2 years ago
#41 - Padding tokens represented differently in different parts of the code
Issue -
State: open - Opened by asmadotgh over 2 years ago
#40 - More Training Information on Reacher
Issue -
State: open - Opened by ivo-1 over 2 years ago
#39 - Problem Creating Deterministic Action Selection
Issue -
State: open - Opened by carsonsmith87 over 2 years ago
- 1 comment
#38 - Confusion over shape of returns_to_go in get_batch
Issue -
State: open - Opened by DaveyBiggers over 2 years ago
#37 - Some problems after reading the paper and code
Issue -
State: open - Opened by CodingNovice7 over 2 years ago
#36 - Padding / attention_mask questions
Issue -
State: closed - Opened by DaveyBiggers over 2 years ago
- 2 comments
#35 - Where are the weights for the trained models?
Issue -
State: closed - Opened by simoninithomas over 2 years ago
- 1 comment
#34 - Error when loading fixed replay buffer
Issue -
State: open - Opened by dgjung0220 over 2 years ago
- 12 comments
#33 - Minor bug that removes best performing trajectory in gym experiments
Issue -
State: closed - Opened by micahcarroll over 2 years ago
- 1 comment
#32 - About parameters in code
Issue -
State: closed - Opened by CodingNovice7 almost 3 years ago
- 1 comment
#31 - TypeError on trajectory_gpt2.py
Issue -
State: closed - Opened by dljzx almost 3 years ago
- 2 comments
#30 - Potential bug: Attention mask allows access to future tokens?
Issue -
State: closed - Opened by donthomasitos almost 3 years ago
- 5 comments
#29 - Can you add reacher dataset
Issue -
State: open - Opened by jsun57 almost 3 years ago
- 1 comment
#28 - about graph experiment
Issue -
State: open - Opened by ysymyth almost 3 years ago
#27 - batch sampling: only last tokens?
Issue -
State: closed - Opened by Howuhh almost 3 years ago
- 1 comment
#26 - Question: is it possible to use the same Decision Transformer for new training trajectories generation?
Issue -
State: closed - Opened by danielgafni almost 3 years ago
- 2 comments
#25 - State and Return preds input
Issue -
State: closed - Opened by backpropper almost 3 years ago
- 1 comment
#24 - undestanding use of rewards
Issue -
State: closed - Opened by jeweinb about 3 years ago
- 2 comments
#23 - Return-to-go conditioning on Atari
Issue -
State: closed - Opened by geekyutao about 3 years ago
- 2 comments
#22 - Possible misalignment in calculating rtg in Atari
Issue -
State: closed - Opened by geekyutao about 3 years ago
- 1 comment
#21 - Regarding atari breakout results
Issue -
State: closed - Opened by dido1998 about 3 years ago
- 1 comment
#20 - aligning action embeddings to other embeddings at line 237
Issue -
State: closed - Opened by loct824 about 3 years ago
- 3 comments
#19 - Timesteps Shape
Issue -
State: closed - Opened by MrShininnnnn about 3 years ago
- 1 comment
#18 - Add medium-expert dataset
Pull Request -
State: open - Opened by ChenDRAG about 3 years ago
#17 - difference between two GPT models used in this repo?
Issue -
State: closed - Opened by ChenDRAG about 3 years ago
- 1 comment
#16 - how to get the score of an expert policy and some other details
Issue -
State: closed - Opened by TianhongDai about 3 years ago
- 1 comment
#15 - Misalignment supervision when predicting succesor state
Issue -
State: closed - Opened by geekyutao about 3 years ago
- 1 comment
#14 - Application to multi-agent environment
Issue -
State: closed - Opened by skull8888888 about 3 years ago
- 1 comment
#13 - RuntimeError
Issue -
State: open - Opened by ximinng about 3 years ago
- 1 comment
#12 - undefined name 'returns
Issue -
State: closed - Opened by l3str4nge about 3 years ago
- 2 comments
#11 - Just a little correction. A parameter changed in script.
Issue -
State: closed - Opened by shogi880 over 3 years ago
- 1 comment
#10 - remove shadowing
Pull Request -
State: closed - Opened by enosair over 3 years ago
- 1 comment
#9 - Getting killed before loading any data
Issue -
State: closed - Opened by YashButala over 3 years ago
- 1 comment
#8 - After loading 50 trajectories, the terminal shows `killed`
Issue -
State: closed - Opened by SunHaoOne over 3 years ago
- 8 comments
#7 - Citation request
Issue -
State: closed - Opened by 222464 over 3 years ago
- 2 comments
#6 - Is this a bug?
Issue -
State: closed - Opened by yuanmao over 3 years ago
- 1 comment
#5 - state and action prediction
Issue -
State: closed - Opened by mehdimashayekhi over 3 years ago
- 4 comments
#2 - Do both BC and DT fit the training data well?
Issue -
State: closed - Opened by w-hc over 3 years ago
- 1 comment
#1 - You forgot to include wandb
Pull Request -
State: closed - Opened by zitterbewegung over 3 years ago