Spaces:
Sleeping
Sleeping
File size: 3,375 Bytes
a168405 38c87f8 a168405 38c87f8 a168405 38c87f8 a168405 38c87f8 a168405 38c87f8 a168405 38c87f8 a168405 38c87f8 a168405 38c87f8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 | import os
import sys
from pathlib import Path
from huggingface_hub import HfApi, login
HF_TOKEN = os.environ.get("HF_TOKEN", "YOUR_HF_TOKEN")
MODEL_REPO = "Chirag0123/shiftlog-gym-qwen-memory-policy"
ADAPTER_DIR = Path("outputs/grpo-stagec")
MODEL_CARD_PATH = Path("model_card.md")
if HF_TOKEN == "YOUR_HF_TOKEN":
print("⚠️ Please set your HF_TOKEN in the script or environment.")
exit(1)
login(token=HF_TOKEN)
api = HfApi(token=HF_TOKEN)
print(f"Checking repo {MODEL_REPO}...")
api.create_repo(repo_id=MODEL_REPO, repo_type="model", exist_ok=True)
def upload_model_card():
if not MODEL_CARD_PATH.exists():
print(f"ERROR: {MODEL_CARD_PATH} does not exist.")
sys.exit(1)
print(f"Uploading model card from {MODEL_CARD_PATH}...")
api.upload_file(
path_or_fileobj=str(MODEL_CARD_PATH),
path_in_repo="README.md",
repo_id=MODEL_REPO,
repo_type="model",
)
print("✅ README.md uploaded")
def upload_adapter_files():
global ADAPTER_DIR
if not ADAPTER_DIR.exists():
print(f"ERROR: {ADAPTER_DIR} does not exist.")
print("Adapter was not saved. Check training pipeline.")
for alt in ["outputs/stage-c", "outputs/lora", "checkpoints/step_300", "checkpoints/final", "shiftlog-model-merged"]:
if Path(alt).exists():
print(f"Found alternate path: {alt}")
ADAPTER_DIR = Path(alt)
break
else:
print("No adapter directories found. Ensure training has run.")
sys.exit(1)
files = list(ADAPTER_DIR.glob("*"))
print(f"Found {len(files)} files in {ADAPTER_DIR}:")
for f in files:
print(f" {f.name} ({f.stat().st_size / 1024:.1f} KB)")
for f in ADAPTER_DIR.glob("*"):
if f.is_file() and not f.name.startswith("."):
print(f"Uploading {f.name}...")
api.upload_file(
path_or_fileobj=str(f),
path_in_repo=f.name,
repo_id=MODEL_REPO,
repo_type="model",
)
print(f" ✅ {f.name} uploaded")
adapter_config_path = ADAPTER_DIR / "adapter_config.json"
if not adapter_config_path.exists():
import json
minimal_config = {
"base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
"bias": "none",
"fan_in_fan_out": False,
"inference_mode": True,
"init_lora_weights": True,
"lora_alpha": 16,
"lora_dropout": 0.0,
"modules_to_save": None,
"peft_type": "LORA",
"r": 8,
"target_modules": ["q_proj", "v_proj", "k_proj", "o_proj", "gate_proj", "up_proj", "down_proj"],
"task_type": "CAUSAL_LM"
}
api.upload_file(
path_or_fileobj=json.dumps(minimal_config, indent=2).encode(),
path_in_repo="adapter_config.json",
repo_id=MODEL_REPO,
repo_type="model",
)
print("✅ Created and uploaded minimal adapter_config.json")
if __name__ == "__main__":
mode = sys.argv[1].strip().lower() if len(sys.argv) > 1 else "all"
if mode in {"all", "adapter"}:
upload_adapter_files()
if mode in {"all", "card"}:
upload_model_card()
print(f"\n✅ Done. Check: https://huggingface.co/{MODEL_REPO}")
|