-
Do you have a plan to support JetMoE model (https://github.com/myshell-ai/JetMoE) that very effective to reduce computational cost in inference in litgpt?
-
### Description
After delivering my session in DevNexus, two devs approached me asking on how to compile a Quarkus + DJL library into native, as DJL library already supports GraalVM compilation https…
-
Error occurred when executing Yoloworld_ESAM_Zho:
cannot import name 'packaging' from 'pkg_resources' (E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\pkg_resources\__init__.py)
…
-
Add API support so that you can use models from other devices
-
### Expected Behavior
Graphics card: Nvidia P40
Model: Flux Schnell - float8_e4m3fn format
After commit 8115d8cce97a3edaaad8b08b45ab37c6782e1cb4 ComfyUI takes about 3x the time to generate an i…
-
When I use https://huggingface.co/spaces/myshell-ai/OpenVoiceV2 for inference, I find that the performance is inconsistent with the reference audio.
Text: 该努力我也努力过了,该尝试的我都尝试过了!
Reference audio: ht…
-
Dear community,
I'm excited to share Colossal-AI, a deep learning framework for training and inference tasks in large language models (LLMs).
Colossal-AI stands out for its exceptional speed, su…
-
### (v17) Custom inference fails when inputs and outputs of the body have mismatched shapes.
The following code should generate a loop that produces a tensor `[1 1 1 1 1]` by appending `1` in each …
-
**System information**
- Android Device information :
samsung/o1qzhx/o1q:12/SP1A.210812.016/G9910ZHU2CVG2:user/release-keys
- TensorFlow Lite in Play Services SDK version :
'com.google.androi…
-
- https://ai.google.dev/edge/mediapipe/solutions/genai/llm_inference
- https://pub.dev/packages/mediapipe_genai
- https://github.com/google/flutter-mediapipe/discussions/62