Skip to content

Actions: AtomicBot-ai/dflash

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
17 workflow runs
17 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Enhance Gemma 4 tool call and reasoning handling in inference server Build & Release DFlash MLX Server (macOS ARM64) #17: Commit 1936b7a pushed by Vect0rM
3m 53s main
main
3m 53s
Refactor inference thread initialization to ensure MLX operations occ… Build & Release DFlash MLX Server (macOS ARM64) #16: Commit 61b57ba pushed by Vect0rM
3m 26s main
main
3m 26s
Replace thread-local GPU stream with default stream in mlx_lm.generat… Build & Release DFlash MLX Server (macOS ARM64) #15: Commit 949c1c4 pushed by Vect0rM
2m 56s main
main
2m 56s
Pin MLX inference work to a single thread by introducing a ThreadPool… Build & Release DFlash MLX Server (macOS ARM64) #14: Commit 7ac8b49 pushed by Vect0rM
2m 47s main
main
2m 47s
Implement custom sanitization for Gemma 4 model weights to drop orpha… Build & Release DFlash MLX Server (macOS ARM64) #13: Commit 2c70d44 pushed by Vect0rM
3m 29s main
main
3m 29s
Update mlx-lm dependency version in pyproject.toml to 0.31.3 for impr… Build & Release DFlash MLX Server (macOS ARM64) #12: Commit fe12300 pushed by Vect0rM
2m 49s main
main
2m 49s
Fix token accumulation in _do_generate function by appending generate… Build & Release DFlash MLX Server (macOS ARM64) #11: Commit 1ed2330 pushed by Vect0rM
3m 11s main
main
3m 11s
Enhance _extract_text function to support chat_template_kwargs Build & Release DFlash MLX Server (macOS ARM64) #10: Commit 89d59f0 pushed by Vect0rM
3m 45s main
main
3m 45s
Implement message normalization for chat templates in inference server Build & Release DFlash MLX Server (macOS ARM64) #9: Commit f486034 pushed by Vect0rM
2m 40s main
main
2m 40s
Enhance tool call parsing and logging in inference server Build & Release DFlash MLX Server (macOS ARM64) #8: Commit a6764b6 pushed by Vect0rM
3m 16s main
main
3m 16s
Enhance tool call handling in inference server Build & Release DFlash MLX Server (macOS ARM64) #7: Commit 177009c pushed by Vect0rM
3m 18s main
main
3m 18s
Update max_tokens handling in chat completions to support new paramet… Build & Release DFlash MLX Server (macOS ARM64) #6: Commit b742c32 pushed by Vect0rM
2m 57s main
main
2m 57s
Add handling for Qwen3-style templates in generation prompt Build & Release DFlash MLX Server (macOS ARM64) #5: Commit a53b897 pushed by Vect0rM
3m 1s main
main
3m 1s
Enhance draft model loading and response handling Build & Release DFlash MLX Server (macOS ARM64) #4: Commit 18d013e pushed by Vect0rM
3m 37s main
main
3m 37s
fix: run MLX generation in thread pool + add server integration tests Build & Release DFlash MLX Server (macOS ARM64) #3: Commit 645586f pushed by Vect0rM
3m 33s main
main
3m 33s
Refactor model loading logic to handle file paths correctly and impro… Build & Release DFlash MLX Server (macOS ARM64) #2: Commit 643d418 pushed by Vect0rM
3m 11s main
main
3m 11s
Add server dependencies for MLX backend in pyproject.toml Build & Release DFlash MLX Server (macOS ARM64) #1: Commit e14f921 pushed by Vect0rM
3m 41s main
main
3m 41s