Spaces:

agentDebugger
/

AgentDebugger-training-v3

Sleeping

App Files Files Community

AgentDebugger-training-v3

Commit History

Jul 8

11:12
fix: remove Dockerfile - unused in sdk:gradio mode, was blocking build⁠

shank +1: Cursor

182d3be
11:05
fix: remove gradio from requirements.txt (HF Spaces pre-installs it, causes pip conflict)⁠

shank +1: Cursor

76ccb75
10:36
Change SDK back to gradio

shank

2b5f690
10:29
fix: restore Dockerfile and use sdk: docker to comply with hackathon health checks

shank

75b6d51
09:24
fix: remove invalid sdk_version from frontmatter

shank

bc780b2
08:20
chore: Rename Dockerfile to fix HF Space deploy, and clean up AI slop

shank

13f7b8d
07:58
fix: move css to gr.Blocks() for Gradio 5.x compat

shank

2e70173
07:24
fix: add sdk_version to Space config to fix scheduling failure

shank

7e7bcac
06:13
update: updated readme

shank

f984ce9
06:09
Updated readme

shank

de4232f
05:36
Update curves and add gitignores

shank

0b1d886
05:02
Flatten markdown block backgrounds

shank

154642f
04:54
Polish UI header and update github links

shank

b391cd8
04:43
Fix UI reward breakdown mapping for offline cache trajectories

shank

bec3d00
04:35
Polish Gradio CSS styling and add base Qwen baseline to leaderboard

shank

9f6368b

Jul 7

07:32
Update evaluation results, fix Gradio compatibility, and resolve sandbox execution path on macOS

shank

0a80c48

May 19

07:03
Add --hub-resume and --step-offset to allow resuming directly from HF Hub PEFT checkpoints

shank

8bd8552

May 18

11:03
Revert speed optimizations to prioritize real-world model quality

shank

cf25957
10:58
Optimize training speed for T4 GPUs

shank

2c50d8a
10:32
Fix HF_TOKEN scope issue for CheckpointPushCallback

shank

ca28017
10:15
Made changes to train_grpo

shank

a07dc7f
08:36
Fix links to match correct usernames/orgs

shank

3293f97
08:30
Fix checkpoint persistence, add leaderboard and update HF links

shank

e160aa1

Apr 26

11:18
fix: use GitHub raw URLs for images so README renders on HF Space

shank

3eb8edc
11:16
Added blog post

shank

c02b65b
11:14
Added blog post

shank

eacdf84
11:08
Added readme again

shank

3165754
11:08
Added readme

shank

f6f33cf
11:06
Update: Added final imporvements for hackathon

shank

713f336
11:02
chore: clean up local dev files and temporary virtual environments

shank

59986c5
11:01
chore: clean up local dev files and temporary virtual environments

shank

374c6cc
06:12
Update: Triggering the full run

shank

75cd77b
05:44
Fix: batch%num_generations math

shank

2b499e7
05:35
Cuda returns false fixed

shank

b8172c5
05:26
COMPUTE_DRIVE fix

shank

77156dd
05:20
Fix: Removed BitsandBytes

shank

bdec91d
05:09
Fix: Fixed dependancy issues

shank

db12eaa
05:07
Revert "Fix: Dockerfile"⁠

shank

dc7eb3f
05:03
Fix: Dockerfile

shank

5dcd156
04:54
Fix: Fixed again again

shank

accb271
04:48
Fix: Fixed again

shank

9864e61
04:35
Fix: Fixing Again

shank

6747185
04:30
Fix: Fixing

shank

18b4e8a
04:22
Fix: Trying to fix dependency issues

shank

024f3c7
04:09
Fix: Fixed file

shank

cb09ef1
03:55
fix: serialize bug_metadata as JSON to fix pyarrow mixed-type error

shank

4668456
03:40
fix: upgrade bitsandbytes>=0.49.0 (triton.ops), switch to Qwen2.5-Coder-3B

shank

a2fa47a
03:25
fix: torch at build time, remove mergekit (conflicts accelerate/peft/trl)

shank

2bfaf77
03:14
fix: empty requirements.txt, install training deps at runtime

shank

5d0b2d4
03:02
fix: remove wandb - click conflict with gradio causes resolution-too-deep

shank

2005cd2

Commit History

Jul 8

fix: remove Dockerfile - unused in sdk:gradio mode, was blocking build⁠

fix: remove gradio from requirements.txt (HF Spaces pre-installs it, causes pip conflict)⁠

Change SDK back to gradio

fix: restore Dockerfile and use sdk: docker to comply with hackathon health checks

fix: remove invalid sdk_version from frontmatter

chore: Rename Dockerfile to fix HF Space deploy, and clean up AI slop

fix: move css to gr.Blocks() for Gradio 5.x compat

fix: add sdk_version to Space config to fix scheduling failure

update: updated readme

Updated readme

Update curves and add gitignores

Flatten markdown block backgrounds

Polish UI header and update github links

Fix UI reward breakdown mapping for offline cache trajectories

Polish Gradio CSS styling and add base Qwen baseline to leaderboard

Jul 7

Update evaluation results, fix Gradio compatibility, and resolve sandbox execution path on macOS

May 19

Add --hub-resume and --step-offset to allow resuming directly from HF Hub PEFT checkpoints

May 18

Revert speed optimizations to prioritize real-world model quality

Optimize training speed for T4 GPUs

Fix HF_TOKEN scope issue for CheckpointPushCallback

Made changes to train_grpo

Fix links to match correct usernames/orgs

Fix checkpoint persistence, add leaderboard and update HF links

Apr 26

fix: use GitHub raw URLs for images so README renders on HF Space

Added blog post

Added blog post

Added readme again

Added readme

Update: Added final imporvements for hackathon

chore: clean up local dev files and temporary virtual environments

chore: clean up local dev files and temporary virtual environments

Update: Triggering the full run

Fix: batch%num_generations math

Cuda returns false fixed

COMPUTE_DRIVE fix

Fix: Removed BitsandBytes

Fix: Fixed dependancy issues

Revert "Fix: Dockerfile"⁠

Fix: Dockerfile

Fix: Fixed again again

Fix: Fixed again

Fix: Fixing Again

Fix: Fixing

Fix: Trying to fix dependency issues

Fix: Fixed file

fix: serialize bug_metadata as JSON to fix pyarrow mixed-type error

fix: upgrade bitsandbytes>=0.49.0 (triton.ops), switch to Qwen2.5-Coder-3B

fix: torch at build time, remove mergekit (conflicts accelerate/peft/trl)

fix: empty requirements.txt, install training deps at runtime

fix: remove wandb - click conflict with gradio causes resolution-too-deep