Update app.py

app.py

@@ -283,13 +283,7 @@ async def upload_images(
 
 
 
-def auto_run_lora_from_repo(folder_path: str):
-    try:
-        FOLDER_IN_REPO = folder_path
-        print(f"🚀 Auto-run triggered for: {FOLDER_IN_REPO}")
-        # Do your training etc...
-    except Exception as e:
-        print(f"❌ auto_run_lora_from_repo failed: {str(e)}")
+
 
 
 
@@ -507,132 +501,22 @@ def recursive_update(d, u):
 
 
 
-@spaces.GPU(duration=50)
-def start_training0(
-    lora_name,
-    concept_sentence,
-    steps,
-    lr,
-    rank,
-    model_to_train,
-    low_vram,
-    dataset_folder,
-    sample_1,
-    sample_2,
-    sample_3,
-    use_more_advanced_options,
-    more_advanced_options,
-):
-    try:
-        user = whoami()
-        username = user.get("name", "anonymous")
-        push_to_hub = True
-    except:
-        username = "anonymous"
-        push_to_hub = False
-
-    slugged_lora_name = lora_name.replace(" ", "_").lower()
-    print(username)
-
-    # Load base config
-    config = {
-        "job": "extension",
-        "config": {
-            "name": slugged_lora_name,
-            "process": [
-                {   "type":"sd_trainer",
-                    "model": {
-                        "low_vram": low_vram,
-                        "is_flux": True,
-                        "quantize": True,
-                        "name_or_path": "black-forest-labs/FLUX.1-dev"
-                    },
-                    "network": {
-                        "linear": rank,
-                        "linear_alpha": rank,
-                        "type": "lora"
-                    },
-                    "train": {
-                        "steps": steps,
-                        "lr": lr,
-                        "skip_first_sample": True,
-                        "batch_size": 1,
-                        "dtype": "bf16",
-                        "gradient_accumulation_steps": 1,
-                        "gradient_checkpointing": True,
-                        "noise_scheduler": "flowmatch",
-                        "optimizer": "adamw8bit",
-                        "ema_config": {
-                            "use_ema": True,
-                            "ema_decay": 0.99
-                        }
-                    },
-                    "datasets": [
-                        {"folder_path": dataset_folder}
-                    ],
-                    "save": {
-                        "dtype": "float16",
-                        "save_every": 10000,
-                        "push_to_hub": push_to_hub,
-                        "hf_repo_id": f"{username}/{slugged_lora_name}",
-                        "hf_private": True,
-                        "max_step_saves_to_keep": 4
-                    },
-                    "sample": {
-                        "guidance_scale": 3.5,
-                        "sample_every": steps,
-                        "sample_steps": 28,
-                        "width": 1024,
-                        "height": 1024,
-                        "walk_seed": True,
-                        "seed": 42,
-                        "sampler": "flowmatch",
-                        "prompts": [p for p in [sample_1, sample_2, sample_3] if p]
-                    },
-                    "trigger_word": concept_sentence
-                }
-            ]
-        }
-    }
-
-    # Apply advanced YAML overrides if any
-    # if use_more_advanced_options and more_advanced_options:
-    #     advanced_config = yaml.safe_load(more_advanced_options)
-    #     config["config"]["process"][0] = recursive_update(config["config"]["process"][0], advanced_config)
-
-    # Save YAML config
-    os.makedirs("/tmp/tmp_configs", exist_ok=True)
-    config_path = f"/tmp/tmp_configs/{uuid.uuid4()}_{slugged_lora_name}.yaml"
-    with open(config_path, "w") as f:
-        yaml.dump(config, f)
-   
-    
-    print(config_path)
-    # Simulate training
-    job = get_job(config_path)
-    job.run()
-    job.cleanup()
-    print(f"[INFO] Starting training with config: {config_path}")
-    print(json.dumps(config, indent=2))
-    return f"Training started successfully with config: {config_path}"
-
-# ========== MAIN ENDPOINT ==========
-
-
 
 
 
 
 
 
+        
 
 
 @app.post("/train-from-hf")
-def auto_run_lora_from_repo():
+def auto_run_lora_from_repo(folder_path: str):
     try:
+        print("Training has kickstarted")
         # ✅ Static or dynamic config
         REPO_ID = "rahul7star/ohamlab"
-        FOLDER_IN_REPO = "filter-demo/upload_20250708_041329_9c5c81"
+        FOLDER_IN_REPO = folder_path
         CONCEPT_SENTENCE = "ohamlab style"
         LORA_NAME = "ohami_filter_autorun"