Could not connect to the generation server error for I2V and T2V #83
Description
I get the error: Generation failed. Could not connect to the generation server. Make sure the backend is running. when trying to do I2V or T2V.
Then immediately after it just shows: Reconnecting. The backend process stopped unexpectedly. Attempting to restart...
Things I have tried:
- Switching encoders to use the API
- Running the app as Admin
- Changing image resolutions
- Turning on/off prompt enhancement
Text to image works fine, as well as generating videos using the API.
Full log:
2026-04-02 21:17:00,675 - INFO - [Electron] Session log file: C:\Users\tyler\AppData\Local\LTXDesktop\logs\session_2026-04-03_02-17-00_unknown.log
2026-04-02 21:17:00,722 - INFO - [Electron] [icon] Loading app icon from: C:\Users\tyler\AppData\Local\Programs\LTX Desktop\resources\icon.ico | exists: false
2026-04-02 21:17:00,843 - INFO - [Renderer] Projects saved: 1
2026-04-02 21:17:00,844 - INFO - [Renderer] Starting Python backend...
2026-04-02 21:17:00,845 - INFO - [Electron] Using bundled Python: C:\Users\tyler\AppData\Local\LTXDesktop\python\python.exe
2026-04-02 21:17:00,845 - INFO - [Electron] Starting Python backend: C:\Users\tyler\AppData\Local\LTXDesktop\python\python.exe C:\Users\tyler\AppData\Local\Programs\LTX Desktop\resources\backend\ltx2_server.py
2026-04-02 21:17:01,428 - ERROR - [Backend] C:\Users\tyler\AppData\Local\LTXDesktop\python\Lib\site-packages\torch\cuda_init_.py:65: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
2026-04-02 21:17:01,429 - ERROR - [Backend] import pynvml # type: ignore[import]
2026-04-02 21:17:02,593 - INFO - [Backend] INFO:main:SageAttention enabled - attention operations will be faster
2026-04-02 21:17:02,606 - INFO - [Backend] INFO:main:Models directory: C:\Users\tyler\AppData\Local\LTXDesktop\models
2026-04-02 21:17:02,641 - INFO - [Backend] INFO:main:Runtime policy force_api_generations=False (system=Windows cuda_available=True vram_gb=23)
2026-04-02 21:17:05,740 - INFO - [Backend] INFO:handlers.settings_handler:Settings loaded from C:\Users\tyler\AppData\Local\LTXDesktop\settings.json
2026-04-02 21:17:05,759 - INFO - [Electron] Checking for update...
2026-04-02 21:17:05,769 - INFO - [Backend] INFO:main:============================================================
2026-04-02 21:17:05,769 - INFO - [Backend] INFO:main:LTX-2 Video Generation Server (FastAPI + Uvicorn)
2026-04-02 21:17:05,790 - INFO - [Backend] INFO:main:Platform: Windows (AMD64)
2026-04-02 21:17:05,791 - INFO - [Backend] INFO:main:Device: cuda | Dtype: torch.bfloat16
2026-04-02 21:17:05,791 - INFO - [Backend] INFO:main:GPU: NVIDIA GeForce RTX 4090 | VRAM: 23 GB
2026-04-02 21:17:05,791 - INFO - [Backend] INFO:main:SageAttention: enabled
2026-04-02 21:17:05,791 - INFO - [Backend] INFO:main:Python: 3.13.12 | Torch: 2.10.0+cu128
2026-04-02 21:17:05,791 - INFO - [Backend] INFO:main:============================================================
2026-04-02 21:17:05,816 - INFO - [Backend] Started server process [33244]
2026-04-02 21:17:05,816 - INFO - [Backend] Waiting for application startup.
2026-04-02 21:17:05,816 - INFO - [Backend] Application startup complete.
2026-04-02 21:17:05,816 - INFO - [Backend] Server running on http://127.0.0.1:55353
2026-04-02 21:17:05,817 - INFO - [Renderer] Checking backend health...
2026-04-02 21:17:05,817 - INFO - [Renderer] Python backend started successfully
2026-04-02 21:17:05,853 - INFO - [Renderer] Backend health: {"status":"ok","models_loaded":false,"active_model":null,"gpu_info":{"name":"NVIDIA GeForce RTX 4090","vram":24564,"vramUsed":2601},"sage_attention":true,"models_status":[{"id":"fast","name":"LTX-2 Fast (Distilled)","loaded":false,"downloaded":true}]}
2026-04-02 21:17:05,994 - INFO - [Backend] INFO:_routes.settings:Applied settings patch (changed=none)
2026-04-02 21:17:22,497 - INFO - [Backend] INFO:handlers.video_generation_handler:Resolution 540p - using fast pipeline
2026-04-02 21:17:22,515 - INFO - [Backend] INFO:handlers.video_generation_handler:Image: C:\Users\tyler\Desktop\TestLTX4.jpg -> 544x960
2026-04-02 21:17:22,516 - INFO - [Backend] INFO:services.text_encoder.ltx_text_encoder:Installed PromptEncoder.init patch for None gemma_root
2026-04-02 21:17:22,516 - INFO - [Backend] INFO:services.text_encoder.ltx_text_encoder:Installed PromptEncoder API embeddings patch
2026-04-02 21:17:22,516 - INFO - [Backend] WARNING:services.text_encoder.ltx_text_encoder:Failed to patch cleanup_memory for module ltx_pipelines.retake_pipeline
2026-04-02 21:17:22,516 - INFO - [Backend] Traceback (most recent call last):
2026-04-02 21:17:22,517 - INFO - [Backend] File "C:\Users\tyler\AppData\Local\Programs\LTX Desktop\resources\backend\services\text_encoder\ltx_text_encoder.py", line 156, in install_cleanup_memory_patch
2026-04-02 21:17:22,517 - INFO - [Backend] module = import(module_name, fromlist=["cleanup_memory"])
2026-04-02 21:17:22,517 - INFO - [Backend] ModuleNotFoundError: No module named 'ltx_pipelines.retake_pipeline'
2026-04-02 21:17:22,517 - INFO - [Backend] INFO:services.text_encoder.ltx_text_encoder:Installed cleanup_memory patch
2026-04-02 21:17:22,517 - INFO - [Backend] INFO:handlers.video_generation_handler:[i2v] Generation started (model=fast, 544x960, 121 frames, 24 fps)
2026-04-02 21:17:22,517 - INFO - [Backend] INFO:services.text_encoder.ltx_text_encoder:Installed PromptEncoder.init patch for None gemma_root
2026-04-02 21:17:22,517 - INFO - [Backend] INFO:handlers.video_generation_handler:[i2v] Pipeline load: 0.00s
2026-04-02 21:17:28,209 - INFO - [Backend] INFO:services.text_encoder.ltx_text_encoder:Text encoded via API in 5.6s
2026-04-02 21:17:28,210 - INFO - [Backend] INFO:handlers.video_generation_handler:[i2v] Text encoding (api): 5.65s
2026-04-02 21:17:28,329 - INFO - [Electron] Python backend exited with code 3221225477
2026-04-02 21:17:28,329 - INFO - [Electron] Using bundled Python: C:\Users\tyler\AppData\Local\LTXDesktop\python\python.exe
2026-04-02 21:17:28,329 - INFO - [Electron] Starting Python backend: C:\Users\tyler\AppData\Local\LTXDesktop\python\python.exe C:\Users\tyler\AppData\Local\Programs\LTX Desktop\resources\backend\ltx2_server.py
2026-04-02 21:17:28,869 - ERROR - [Backend] C:\Users\tyler\AppData\Local\LTXDesktop\python\Lib\site-packages\torch\cuda_init.py:65: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
2026-04-02 21:17:28,869 - ERROR - [Backend] import pynvml # type: ignore[import]
2026-04-02 21:17:29,974 - INFO - [Backend] INFO:main:SageAttention enabled - attention operations will be faster
2026-04-02 21:17:29,985 - INFO - [Backend] INFO:main:Models directory: C:\Users\tyler\AppData\Local\LTXDesktop\models
2026-04-02 21:17:30,019 - INFO - [Backend] INFO:main:Runtime policy force_api_generations=False (system=Windows cuda_available=True vram_gb=23)
2026-04-02 21:17:33,103 - INFO - [Backend] INFO:handlers.settings_handler:Settings loaded from C:\Users\tyler\AppData\Local\LTXDesktop\settings.json
2026-04-02 21:17:33,130 - INFO - [Backend] INFO:main:============================================================
2026-04-02 21:17:33,130 - INFO - [Backend] INFO:main:LTX-2 Video Generation Server (FastAPI + Uvicorn)
2026-04-02 21:17:33,151 - INFO - [Backend] INFO:main:Platform: Windows (AMD64)
2026-04-02 21:17:33,151 - INFO - [Backend] INFO:main:Device: cuda | Dtype: torch.bfloat16
2026-04-02 21:17:33,151 - INFO - [Backend] INFO:main:GPU: NVIDIA GeForce RTX 4090 | VRAM: 23 GB
2026-04-02 21:17:33,151 - INFO - [Backend] INFO:main:SageAttention: enabled
2026-04-02 21:17:33,152 - INFO - [Backend] INFO:main:Python: 3.13.12 | Torch: 2.10.0+cu128
2026-04-02 21:17:33,152 - INFO - [Backend] INFO:main:============================================================
2026-04-02 21:17:33,176 - INFO - [Backend] Started server process [37352]
2026-04-02 21:17:33,176 - INFO - [Backend] Waiting for application startup.
2026-04-02 21:17:33,177 - INFO - [Backend] Application startup complete.
2026-04-02 21:17:33,177 - INFO - [Backend] Server running on http://127.0.0.1:60119
2026-04-02 21:17:33,178 - INFO - [Renderer] Checking backend health...
2026-04-02 21:17:33,198 - INFO - [Renderer] Backend health: {"status":"ok","models_loaded":false,"active_model":null,"gpu_info":{"name":"NVIDIA GeForce RTX 4090","vram":24564,"vramUsed":2708},"sage_attention":true,"models_status":[{"id":"fast","name":"LTX-2 Fast (Distilled)","loaded":false,"downloaded":true}]}
2026-04-02 21:17:33,340 - INFO - [Backend] INFO:_routes.settings:Applied settings patch (changed=none)