Spaces:
Sleeping
Sleeping
Validate checkpoint credentials once; log throughput
#3
by arminfg - opened
app.py
CHANGED
|
@@ -21,6 +21,7 @@ ENV_NAME = os.environ.get("ENV_NAME", "G1JoystickRoughTerrain")
|
|
| 21 |
NUM_TIMESTEPS = int(os.environ.get("NUM_TIMESTEPS", 200_000_000))
|
| 22 |
SEED = int(os.environ.get("SEED", 0))
|
| 23 |
HF_REPO = os.environ.get("HF_REPO", "").strip()
|
|
|
|
| 24 |
AUTO_START = os.environ.get("AUTO_START", "1") == "1"
|
| 25 |
|
| 26 |
# /data exists only when persistent storage is attached; fall back to /tmp.
|
|
@@ -39,18 +40,49 @@ class Stopped(Exception):
|
|
| 39 |
pass
|
| 40 |
|
| 41 |
|
| 42 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
if not HF_REPO:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
return
|
| 45 |
try:
|
| 46 |
from huggingface_hub import HfApi
|
| 47 |
-
api = HfApi()
|
|
|
|
| 48 |
api.create_repo(HF_REPO, repo_type="model", exist_ok=True)
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
except Exception as e: # never let upload kill training
|
| 53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
|
| 56 |
def install_jax_pmap_shims() -> None:
|
|
@@ -127,14 +159,24 @@ def train_worker() -> None:
|
|
| 127 |
f"timesteps={NUM_TIMESTEPS:,}")
|
| 128 |
|
| 129 |
OUT.mkdir(parents=True, exist_ok=True)
|
|
|
|
|
|
|
| 130 |
|
| 131 |
def progress(step, metrics):
|
| 132 |
if STATE["stop"]:
|
| 133 |
raise Stopped()
|
|
|
|
| 134 |
STATE["step"] = int(step)
|
| 135 |
rew = metrics.get("eval/episode_reward", float("nan"))
|
| 136 |
ln = metrics.get("eval/avg_episode_length", float("nan"))
|
| 137 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 138 |
|
| 139 |
def policy_params_fn(step, make_policy, params):
|
| 140 |
p = OUT / f"ckpt_{int(step)}.pkl"
|
|
|
|
| 21 |
NUM_TIMESTEPS = int(os.environ.get("NUM_TIMESTEPS", 200_000_000))
|
| 22 |
SEED = int(os.environ.get("SEED", 0))
|
| 23 |
HF_REPO = os.environ.get("HF_REPO", "").strip()
|
| 24 |
+
HF_TOKEN = os.environ.get("HF_TOKEN", "").strip() or None
|
| 25 |
AUTO_START = os.environ.get("AUTO_START", "1") == "1"
|
| 26 |
|
| 27 |
# /data exists only when persistent storage is attached; fall back to /tmp.
|
|
|
|
| 40 |
pass
|
| 41 |
|
| 42 |
|
| 43 |
+
UPLOAD = {"enabled": False, "fails": 0}
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
def init_upload() -> None:
|
| 47 |
+
"""Validate credentials once, up front, instead of failing on every eval."""
|
| 48 |
if not HF_REPO:
|
| 49 |
+
log(f"HF_REPO unset -- checkpoints stay in {OUT} and are lost on restart")
|
| 50 |
+
return
|
| 51 |
+
if not HF_TOKEN:
|
| 52 |
+
log("HF_TOKEN secret unset -- cannot push checkpoints")
|
| 53 |
return
|
| 54 |
try:
|
| 55 |
from huggingface_hub import HfApi
|
| 56 |
+
api = HfApi(token=HF_TOKEN)
|
| 57 |
+
who = api.whoami().get("name")
|
| 58 |
api.create_repo(HF_REPO, repo_type="model", exist_ok=True)
|
| 59 |
+
UPLOAD["enabled"] = True
|
| 60 |
+
log(f"checkpoints -> https://huggingface.co/{HF_REPO} (as {who})")
|
| 61 |
+
except Exception as e:
|
| 62 |
+
log(f"cannot write to HF_REPO={HF_REPO}: {type(e).__name__}: "
|
| 63 |
+
f"{str(e).splitlines()[0][:200]}")
|
| 64 |
+
log("fix: HF_TOKEN needs write scope on that namespace. A token scoped to "
|
| 65 |
+
"your own user cannot write to an org repo -- point HF_REPO at a "
|
| 66 |
+
"namespace the token owns, e.g. <your-user>/g1-rough-terrain.")
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
def _upload(path: Path) -> None:
|
| 70 |
+
if not UPLOAD["enabled"]:
|
| 71 |
+
return
|
| 72 |
+
try:
|
| 73 |
+
from huggingface_hub import HfApi
|
| 74 |
+
HfApi(token=HF_TOKEN).upload_file(
|
| 75 |
+
path_or_fileobj=str(path), path_in_repo=path.name,
|
| 76 |
+
repo_id=HF_REPO, repo_type="model")
|
| 77 |
+
log(f"uploaded {path.name}")
|
| 78 |
+
UPLOAD["fails"] = 0
|
| 79 |
except Exception as e: # never let upload kill training
|
| 80 |
+
UPLOAD["fails"] += 1
|
| 81 |
+
log(f"upload failed ({type(e).__name__}): {str(e).splitlines()[0][:160]}")
|
| 82 |
+
if UPLOAD["fails"] >= 3:
|
| 83 |
+
UPLOAD["enabled"] = False
|
| 84 |
+
log("disabling uploads after 3 failures; training continues, "
|
| 85 |
+
f"checkpoints remain in {OUT}")
|
| 86 |
|
| 87 |
|
| 88 |
def install_jax_pmap_shims() -> None:
|
|
|
|
| 159 |
f"timesteps={NUM_TIMESTEPS:,}")
|
| 160 |
|
| 161 |
OUT.mkdir(parents=True, exist_ok=True)
|
| 162 |
+
init_upload()
|
| 163 |
+
t_start = [None]
|
| 164 |
|
| 165 |
def progress(step, metrics):
|
| 166 |
if STATE["stop"]:
|
| 167 |
raise Stopped()
|
| 168 |
+
import time
|
| 169 |
STATE["step"] = int(step)
|
| 170 |
rew = metrics.get("eval/episode_reward", float("nan"))
|
| 171 |
ln = metrics.get("eval/avg_episode_length", float("nan"))
|
| 172 |
+
if t_start[0] is None:
|
| 173 |
+
t_start[0] = (time.time(), int(step))
|
| 174 |
+
sps = float("nan")
|
| 175 |
+
else:
|
| 176 |
+
t0, s0 = t_start[0]
|
| 177 |
+
sps = (int(step) - s0) / max(time.time() - t0, 1e-9)
|
| 178 |
+
log(f"step {int(step):>12,} reward {rew:8.2f} ep_len {ln:7.1f} "
|
| 179 |
+
f"{sps:,.0f} steps/s")
|
| 180 |
|
| 181 |
def policy_params_fn(step, make_policy, params):
|
| 182 |
p = OUT / f"ckpt_{int(step)}.pkl"
|