Deploy updated SLM customer-support chatbot

SLM_CService.py

@@ -15,13 +15,11 @@ from peft import PeftModel
 conversation_history = []
 
 # 3) Model + adapter path in your repo (copied into the Space repo root)
-MODEL_DIR = "bitext-qlora-tinyllama"
+MODEL_DIR = "ThomasBasil/bitext-qlora-tinyllama"
 
 # 4) Load tokenizer from local dir
 tokenizer = AutoTokenizer.from_pretrained(
-    MODEL_DIR,
-    use_fast=False,
-    local_files_only=True
+    "ThomasBasil/bitext-qlora-tinyllama", use_fast=False
 )
 tokenizer.pad_token_id = tokenizer.eos_token_id
 tokenizer.padding_side = "left"
@@ -43,7 +41,7 @@ model = unsloth.FastLanguageModel.from_pretrained(
     trust_remote_code=True
 )
 # 5b) Attach your LoRA adapter
-model = PeftModel.from_pretrained(model, MODEL_DIR, local_files_only=True)
+model = PeftModel.from_pretrained(model, "ThomasBasil/bitext-qlora-tinyllama")
 
 # 6) HF text-gen pipeline
 chat_pipe = pipeline(