Youngho Lee

"Save every n steps" in training cause an CUDA out of memory

2

### Describe the bug ``` # Environment GPU: RTX 4090 24GB VRAM Memory: 32GB RAM CPU: i-13700K # Command python server.py --auto-devices --chat --model-menu --gpu-memory 21GiB 21GiB --cpu-memory 24000MiB --load-in-8bit...

Support StableLM

6

I'm curious if you guys will provide StableLM capability on the web-llm? It would be really great if so.

converter does not work with the current ggml

4

Tried to convert `https://huggingface.co/intfloat/e5-large-v2` to ggml with the current `d9f04e609fb7f7e5fb3b20a77d4d685219971009` commit. However, execution of the converted f32, f16, q4_0, and q4_1 models shows the `not enough space in the context's...