Change model to 8B

app.py

@@ -14,7 +14,7 @@ hf_token = os.getenv("HF_TOKEN")
 
 
 # --- Load tokenizer ---
-tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.2-3B", use_fast=True, token=hf_token)
+tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.1-8B", use_fast=True, token=hf_token)
 vocab_size = len(tokenizer)
 eos_token_id = tokenizer.eos_token_id
 mask_token_id = tokenizer.encode('MASK', add_special_tokens=False)[0]
@@ -37,9 +37,9 @@ assistant_marker_ids = tokenizer.encode("Assistant:", add_special_tokens=False)
 def load_model():
     ckpt_path = hf_hub_download(
         repo_id="ruurd/tini_model",
-        filename="diffusion-model.pth",
+        filename="diffusion-model-8B.pth",
         token=os.getenv("HF_TOKEN"),
-        revision="1ffb916dd34f442f87cf06dda74b96f86eaf1d15",
+        # revision="1ffb916dd34f442f87cf06dda74b96f86eaf1d15",
     )
 
     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")