Downloads · 30 days
0
EdwardoSunny/ttt-discover-v3-qwen3-8b-ac2-step3
ttt-discover-v3-qwen3-8b-ac2-step3 is a machine learning model from EdwardoSunny. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
LoRA adapter (rank 32, all-linear) from the upstream-faithful local reproduction of TTT-Discover on autoresearch-bench ac2. Checkpoint saved after training step 2 (0-indexed). Strict upstream eval parity: 1100s hard k…
Downloads · 30 days
0
Access
Public
Updated Jul 9, 2026
Repo size
349 MB
Likes
0
Public
Click a slice to open those files.
.safetensors349 MB · 100%
From the Hugging Face model README
LoRA adapter (rank 32, all-linear) from the upstream-faithful local
reproduction of TTT-Discover on autoresearch-bench ac2. Checkpoint saved
after training step 2 (0-indexed). Strict upstream eval parity:
1100s hard kill, verbatim prompts/entrypoints, group 64x8, T=1.0, kl 0.1.
{
"step": 2,
"progress/batch": 2,
"optim/lr": 4e-05,
"progress/done_frac": 0.06,
"puct/buffer_size": 40,
"puct/sampled_size": 8,
"puct/T": 1024,
"puct/scale_last": 0.4270916158988898,
"puct/buffer_value/mean": 0.8597736028074138,
"puct/buffer_value/std": 0.11637949147278169,
"puct/buffer_value/min": 0.5041471954736918,
"puct/buffer_value/max": 0.9312388113725816,
"puct/buffer_timestep/mean": 0.2,
"puct/buffer_timestep/std": 0.7483314773547883,
"puct/buffer_timestep/min": -1.0,
"puct/buffer_timestep/max": 1.0,
"puct/buffer_construction_len/mean": 2644.275,
"puct/buffer_construction_len/std": 1678.5433415241323,
"puct/buffer_construction_len/min": 1024.0,
"puct/buffer_construction_len/max": 7414.0,
"puct/sampled_value/mean": 0.9251349466756222,
"puct/sampled_value/std": 0.0025808262833416406,
"puct/sampled_value/min": 0.923047908564699,
"puct/sampled_value/max": 0.9312388113725816,
"puct/sampled_timestep/mean": 1.0,
"puct/sampled_timestep/std": 0.0,
"puct/sampled_timestep/min": 1.0,
"puct/sampled_timestep/max": 1.0,
"puct/sampled_construction_len/mean": 2560.0,
"puct/sampled_construction_len/std": 886.8100134752651,
"puct/sampled_construction_len/min": 2048.0,
"puct/sampled_construction_len/max": 4096.0,
"time/sampling": 5513.096382379532,
"env/all/ac_tokens_per_turn": 7896.19140625,
"env/all/ob_tokens_per_turn": 4422.625,
"env/all/turns_per_episode": 1.0,
"env/all/total_episodes": 512,
"env/all/total_turns": 512,
"env/all/total_ac_tokens": 4042850,
"env/all/total_ob_tokens": 2264384,
"env/all/time/sampling_mean": 529.0116261071526,
"env/all/time/sampling_max": 724.7175903320312,
"env/all/time/env_step_mean": 2231.63677679142,
"env/all/time/env_step_max": 4787.595690250397,
"env/all/reward/mean": 0.2422542964770329,
"env/all/reward/max": 0.9353183644517703,
"env/all/reward/min": 0.0,
"env/all/format": 1.0,
"env/all/format/min": 1.0,
"env/all/format/max": 1.0,
"env/all/reward": 0.2422542964770329,
"env/all/correctness": 0.267578125,
"env/all/correctness/min": 0.0,
"env/all/correctness/max": 1.0,
"env/all/raw_score": 0.9053591226002982,
"env/all/raw_score/min": 0.6667019304422238,
"env/all/raw_score/max": 0.9353183644517703,
"env/all/initial_raw_score": 0.9251349466756222,
"env/all/initial_raw_score/min": 0.923047908564699,
"env/all/initial_raw_score/max": 0.9312388113725816,
"env/all/msg": "Program must define a top-level `run_code()` function.",
"env/all/parsed_code": "```python\n# (Other helper functions like _Adam, _objective_and_grad_conv, etc., are as defined previously and are assumed to be available)\n```",
"env/all/time/policy": 529.0116261071526,
"env/all/time/policy/min": 147.14420318603516,
"env/all/time/policy/max": 724.7175903320312,
"env/all/time/env_step": 2231.63677679142,
"env/all/time/env_step/min": 0.005632877349853516,
"env/all/time/env_step/max": 4787.595690250397,
"env/all/time/reward_compute": 4.279427230358124e-07,
"env/all/time/reward_compute/min": 2.2724270820617676e-07,
"env/all/time/reward_compute/max": 8.493661880493164e-07,
"env/all/by_group/frac_mixed": 1.0,
"env/all/by_group/frac_all_good": 0.0,
"env/all/by_group/frac_all_bad": 0.0,
"advantage/mean": 0.027530409395694733,
"advantage/min": -0.8711702823638916,
"advantage/max": 3.353550434112549,
"time/assemble_training_data": 10.116698503494263,
"time/kl_vs_base": 149.13232588768005,
"kl_policy_base": 0.0004241218266543001,
"time/train": 1180.4556849002838,
"time/save_checkpoint": 17.6952121257782,
"time/total": 6872.178877353668
}
[2026-07-09T06:38:34+00:00] job=1812632 node=node-30 ngpu=3 ntrain=1 replicas=2 flash_attn=no
[2026-07-09T06:45:52+00:00] job=1812704 node=node-30 ngpu=3 ntrain=1 replicas=2 flash_attn=no
[2026-07-09T07:00:42+00:00] job=1812735 node=node-1 ngpu=3 ntrain=1 replicas=2 flash_attn=no
[2026-07-09T07:26:33+00:00] job=1812827 node=node-14 ngpu=3 ntrain=1 replicas=2 flash_attn=yes
[2026-07-09T09:21:09+00:00] job=1813131 node=node-1 ngpu=3 ntrain=1 replicas=2 flash_attn=yes
[2026-07-09T14:53:48+00:00] job=1813132 node=node-2 ngpu=6 ntrain=2 replicas=4 flash_attn=yes