Upload app.py with huggingface_hub

app.py

@@ -26,7 +26,9 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStream
 # ── Config ────────────────────────────────────────────────────────────────────
 
 MODEL_ID = os.environ.get("MODEL_ID", "oddadmix/Nawah-Reasoning-v1")
-HF_TOKEN = os.environ.get("HF_TOKEN")  # only needed while the model repo is private
+# Only needed while the model repo is private. HF_TOKEN appears to be reserved by Spaces
+# (a secret set under that name does not reach the container), so MODEL_HF_TOKEN is preferred.
+HF_TOKEN = os.environ.get("MODEL_HF_TOKEN") or os.environ.get("HF_TOKEN")
 
 IM_START, IM_END = "<|im_start|>", "<|im_end|>"
 THINK_OPEN, THINK_CLOSE = "<think>", "</think>"
@@ -36,6 +38,8 @@ MAX_NEW_TOKENS_CAP = 512
 
 torch.set_num_threads(os.cpu_count() or 2)
 
+print("[*] token env vars present:",
+      [k for k in ("MODEL_HF_TOKEN", "HF_TOKEN") if os.environ.get(k)] or "NONE")
 print(f"[*] Loading {MODEL_ID} ...")
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, token=HF_TOKEN)
 model = AutoModelForCausalLM.from_pretrained(MODEL_ID, dtype=torch.float32, token=HF_TOKEN)