47 MB

Ctrl+K

1 contributor

Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters.

1d568ad verified 3 months ago

.gitattributes

1.52 kB
initial commit 3 months ago
0000100_adapters.safetensors

11.8 MB
xet

Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
0000200_adapters.safetensors

11.8 MB
xet

Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
0000300_adapters.safetensors

11.8 MB
xet

Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
LAM_README.txt

451 Bytes
Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
README.md

4.78 kB
Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
adapter_config.json

1.02 kB
Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
adapters.safetensors

11.8 MB
xet

Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago
training_meta.json

411 Bytes
Update to iterative closed-loop LAM: retrained 300 iters on step-by-step data using build_continuation_prompt. Now supports true multi-turn via one-action-per-step + feedback. Updated model card with new usage, evaluation (100% success in 10 iterative tests), and architecture details. Includes retrained adapters. 3 months ago