Upload folder using huggingface_hub
Browse files- al_0.47_g_0.99_102_bf16/action_logprobs.pth.gz +3 -0
- al_0.47_g_0.99_102_bf16/action_probs.tar.gz +3 -0
- al_0.47_g_0.99_102_bf16/checkpoints.tar.gz +3 -0
- al_0.47_g_0.99_102_bf16/config.cfg +46 -0
- al_0.47_g_0.99_102_bf16/config.json +51 -0
- al_0.47_g_0.99_102_bf16/eval.jsonl +0 -0
- al_0.47_g_0.99_102_bf16/eval.log +0 -0
- al_0.47_g_0.99_102_bf16/latest_eval.json +9 -0
- al_0.47_g_0.99_102_bf16/latest_train.json +12 -0
- al_0.47_g_0.99_102_bf16/setup.json +66 -0
- al_0.47_g_0.99_102_bf16/setup.txt +8 -0
- al_0.47_g_0.99_102_bf16/train.jsonl +0 -0
- al_0.47_g_0.99_102_bf16/train.log +0 -0
al_0.47_g_0.99_102_bf16/action_logprobs.pth.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8d9ae9e29136da7ac34c0d3b8a53d7a87d2c5557f58666d282c10f7e47ac087
|
| 3 |
+
size 570178701
|
al_0.47_g_0.99_102_bf16/action_probs.tar.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2705afd39b801f93eb6ae1b6b9e03822f6c3893c15a3f8e8c0437d4f09c2e68
|
| 3 |
+
size 113218474
|
al_0.47_g_0.99_102_bf16/checkpoints.tar.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae6cb71bd7ed046e5aba8ac535bc0fe35c72ead510b0de1522db554f2feba0be
|
| 3 |
+
size 749007159
|
al_0.47_g_0.99_102_bf16/config.cfg
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
rl_action=train
|
| 2 |
+
num_rollout_steps=64
|
| 3 |
+
lr=5e-05
|
| 4 |
+
discount_rate=0.99
|
| 5 |
+
eff_horizon=None
|
| 6 |
+
eval_every=1
|
| 7 |
+
use_wandb=True
|
| 8 |
+
use_hf=True
|
| 9 |
+
use_log=True
|
| 10 |
+
num_total_env_steps=5000000000
|
| 11 |
+
checkpoint=al_0.47_g_0.99_102_bf16
|
| 12 |
+
render_sixel=True
|
| 13 |
+
sixel_loc=(7, 7)
|
| 14 |
+
seed=102
|
| 15 |
+
mask_type=first_episode
|
| 16 |
+
penalize_time=False
|
| 17 |
+
optim=adam
|
| 18 |
+
live_monitor=False
|
| 19 |
+
use_bf16=True
|
| 20 |
+
checkpoint_schedule=0:8
|
| 21 |
+
grad_acc_per_chunk=16
|
| 22 |
+
num_rollout_chunks=1
|
| 23 |
+
cheese_loc=any
|
| 24 |
+
env_layout=open
|
| 25 |
+
alpha=0.47
|
| 26 |
+
env_size=13
|
| 27 |
+
num_levels=9600
|
| 28 |
+
f_str_ckpt=al_{alpha}_g_{discount_rate}_{seed}_bf16
|
| 29 |
+
wandb_project=jaxgmg_3phase_bf16
|
| 30 |
+
ckpt_dir=jaxgmg_3phase_bf16
|
| 31 |
+
duplication_factor=-1
|
| 32 |
+
smoke=False
|
| 33 |
+
num_chains=6
|
| 34 |
+
num_draws=3000
|
| 35 |
+
on_policy=True
|
| 36 |
+
nbeta=3000
|
| 37 |
+
localization=10
|
| 38 |
+
exact_solver_each_draw=False
|
| 39 |
+
llc_optimizer=sgld
|
| 40 |
+
iw_clip_eps=None
|
| 41 |
+
rmsprop_burnin=20
|
| 42 |
+
llc_data_file=llc_scan_open_reinforce.pkl
|
| 43 |
+
llc_checkpoint_index=0
|
| 44 |
+
repo_id=davidquarel/jaxgmg_ckpt_zip
|
| 45 |
+
use_shuffled_checkpoints=0
|
| 46 |
+
force_re_download=False
|
al_0.47_g_0.99_102_bf16/config.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha": 0.47,
|
| 3 |
+
"checkpoint": "al_0.47_g_0.99_102_bf16",
|
| 4 |
+
"checkpoint_schedule": "0:8",
|
| 5 |
+
"cheese_loc": "any",
|
| 6 |
+
"ckpt_dir": "jaxgmg_3phase_bf16",
|
| 7 |
+
"discount_rate": 0.99,
|
| 8 |
+
"duplication_factor": -1,
|
| 9 |
+
"eff_horizon": null,
|
| 10 |
+
"env_layout": "open",
|
| 11 |
+
"env_size": 13,
|
| 12 |
+
"eval_every": 1,
|
| 13 |
+
"exact_solver_each_draw": false,
|
| 14 |
+
"f_str_ckpt": "al_{alpha}_g_{discount_rate}_{seed}_bf16",
|
| 15 |
+
"force_re_download": false,
|
| 16 |
+
"grad_acc_per_chunk": 16,
|
| 17 |
+
"iw_clip_eps": null,
|
| 18 |
+
"live_monitor": false,
|
| 19 |
+
"llc_checkpoint_index": 0,
|
| 20 |
+
"llc_data_file": "llc_scan_open_reinforce.pkl",
|
| 21 |
+
"llc_optimizer": "sgld",
|
| 22 |
+
"localization": 10,
|
| 23 |
+
"lr": 5e-05,
|
| 24 |
+
"mask_type": "first_episode",
|
| 25 |
+
"nbeta": 3000,
|
| 26 |
+
"num_chains": 6,
|
| 27 |
+
"num_draws": 3000,
|
| 28 |
+
"num_levels": 9600,
|
| 29 |
+
"num_rollout_chunks": 1,
|
| 30 |
+
"num_rollout_steps": 64,
|
| 31 |
+
"num_total_env_steps": 5000000000,
|
| 32 |
+
"on_policy": true,
|
| 33 |
+
"optim": "adam",
|
| 34 |
+
"penalize_time": false,
|
| 35 |
+
"render_sixel": true,
|
| 36 |
+
"repo_id": "davidquarel/jaxgmg_ckpt_zip",
|
| 37 |
+
"rl_action": "train",
|
| 38 |
+
"rmsprop_burnin": 20,
|
| 39 |
+
"seed": 102,
|
| 40 |
+
"sixel_loc": [
|
| 41 |
+
7,
|
| 42 |
+
7
|
| 43 |
+
],
|
| 44 |
+
"smoke": false,
|
| 45 |
+
"use_bf16": true,
|
| 46 |
+
"use_hf": true,
|
| 47 |
+
"use_log": true,
|
| 48 |
+
"use_shuffled_checkpoints": 0,
|
| 49 |
+
"use_wandb": true,
|
| 50 |
+
"wandb_project": "jaxgmg_3phase_bf16"
|
| 51 |
+
}
|
al_0.47_g_0.99_102_bf16/eval.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
al_0.47_g_0.99_102_bf16/eval.log
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
al_0.47_g_0.99_102_bf16/latest_eval.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"ts": "2025-10-20T19:07:25.183421Z",
|
| 3 |
+
"loop": 8136,
|
| 4 |
+
"env_steps": 4998758400,
|
| 5 |
+
"exact_value_avg": 0.9147202372550964,
|
| 6 |
+
"regret": 0.010819630697369576,
|
| 7 |
+
"action_probs_image_path": "jaxgmg_3phase_bf16/al_0.47_g_0.99_102_bf16/action_probs/img_8136.png",
|
| 8 |
+
"entropy": 0.3846891522407532
|
| 9 |
+
}
|
al_0.47_g_0.99_102_bf16/latest_train.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"ts": "2025-10-20T19:07:25.781094Z",
|
| 3 |
+
"loop": 8137,
|
| 4 |
+
"env_steps": 4999987199,
|
| 5 |
+
"loss": 0.002683906117454171,
|
| 6 |
+
"avg_return": 0.9147175550460815,
|
| 7 |
+
"regret": 0.010819630697369576,
|
| 8 |
+
"entropy": 0.27988603711128235,
|
| 9 |
+
"lr": 5e-05,
|
| 10 |
+
"discount_rate": 0.99,
|
| 11 |
+
"num_rollout_steps": 64
|
| 12 |
+
}
|
al_0.47_g_0.99_102_bf16/setup.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"config": {
|
| 3 |
+
"rl_action": "train",
|
| 4 |
+
"num_rollout_steps": 64,
|
| 5 |
+
"lr": 5e-05,
|
| 6 |
+
"discount_rate": 0.99,
|
| 7 |
+
"eff_horizon": null,
|
| 8 |
+
"eval_every": 1,
|
| 9 |
+
"use_wandb": true,
|
| 10 |
+
"use_hf": true,
|
| 11 |
+
"use_log": true,
|
| 12 |
+
"num_total_env_steps": 5000000000,
|
| 13 |
+
"checkpoint": "al_0.47_g_0.99_102_bf16",
|
| 14 |
+
"render_sixel": true,
|
| 15 |
+
"sixel_loc": [
|
| 16 |
+
7,
|
| 17 |
+
7
|
| 18 |
+
],
|
| 19 |
+
"seed": 102,
|
| 20 |
+
"mask_type": "first_episode",
|
| 21 |
+
"penalize_time": false,
|
| 22 |
+
"optim": "adam",
|
| 23 |
+
"live_monitor": false,
|
| 24 |
+
"use_bf16": true,
|
| 25 |
+
"checkpoint_schedule": "0:8",
|
| 26 |
+
"grad_acc_per_chunk": 16,
|
| 27 |
+
"num_rollout_chunks": 1,
|
| 28 |
+
"cheese_loc": "any",
|
| 29 |
+
"env_layout": "open",
|
| 30 |
+
"alpha": 0.47,
|
| 31 |
+
"env_size": 13,
|
| 32 |
+
"num_levels": 9600,
|
| 33 |
+
"f_str_ckpt": "al_{alpha}_g_{discount_rate}_{seed}_bf16",
|
| 34 |
+
"wandb_project": "jaxgmg_3phase_bf16",
|
| 35 |
+
"ckpt_dir": "jaxgmg_3phase_bf16",
|
| 36 |
+
"duplication_factor": -1,
|
| 37 |
+
"smoke": false,
|
| 38 |
+
"num_chains": 6,
|
| 39 |
+
"num_draws": 3000,
|
| 40 |
+
"on_policy": true,
|
| 41 |
+
"nbeta": 3000,
|
| 42 |
+
"localization": 10,
|
| 43 |
+
"exact_solver_each_draw": false,
|
| 44 |
+
"llc_optimizer": "sgld",
|
| 45 |
+
"iw_clip_eps": null,
|
| 46 |
+
"rmsprop_burnin": 20,
|
| 47 |
+
"llc_data_file": "llc_scan_open_reinforce.pkl",
|
| 48 |
+
"llc_checkpoint_index": 0,
|
| 49 |
+
"repo_id": "davidquarel/jaxgmg_ckpt_zip",
|
| 50 |
+
"use_shuffled_checkpoints": 0,
|
| 51 |
+
"force_re_download": false
|
| 52 |
+
},
|
| 53 |
+
"setup": {
|
| 54 |
+
"start_time_utc": "2025-10-20T17:47:36.496213Z",
|
| 55 |
+
"seed": 102,
|
| 56 |
+
"device": "cuda",
|
| 57 |
+
"python_version": "3.11.11",
|
| 58 |
+
"torch_version": "2.9.0+cu128",
|
| 59 |
+
"jax_version": "0.6.2",
|
| 60 |
+
"hostname": "0a4b26e7f7a9",
|
| 61 |
+
"platform": "Linux-5.4.0-216-generic-x86_64-with-glibc2.35",
|
| 62 |
+
"git_sha": null,
|
| 63 |
+
"git_branch": null,
|
| 64 |
+
"use_wandb": true
|
| 65 |
+
}
|
| 66 |
+
}
|
al_0.47_g_0.99_102_bf16/setup.txt
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Start: 2025-10-20T17:47:36.496213Z
|
| 2 |
+
Seed: 102
|
| 3 |
+
Device: cuda
|
| 4 |
+
Python: 3.11.11 | Torch: 2.9.0+cu128 | JAX: 0.6.2
|
| 5 |
+
Host: 0a4b26e7f7a9
|
| 6 |
+
Platform: Linux-5.4.0-216-generic-x86_64-with-glibc2.35
|
| 7 |
+
Git: branch=None sha=None
|
| 8 |
+
use_wandb: True
|
al_0.47_g_0.99_102_bf16/train.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
al_0.47_g_0.99_102_bf16/train.log
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|