-
Allocating 1D tensor of type int4 via [`BufferFromHostBuffer`](https://github.com/openxla/xla/blob/41ad400d325243342b11e1d55232b34bbd590b8c/xla/pjrt/cpu/cpu_client.cc#L944) with `byte_strides = std::n…
-
I installed v0.2 on my Fairphone 4 running LineageOS-with-microg 20 (Android 13) and pulled the stock camera app from the latest FP4 firmware. The camera app version is `v2.0.039(06301600-01)`. With y…
-
I use ipex-llm to quantize and push models to hub. But it seems `load_low_bit` expects the model to be locally available and cant take it from huggingface hub.
It would be awesome to allow the mode…
-
### Is your feature request related to a problem? Please describe.
MOSS has beedn added to AutoGPTQ https://github.com/yhyu13/AutoGPTQ/blob/main/auto_gptq/modeling/moss.py So that the community can c…
-
The current FP3 and FP4 fingerprints reparse the SMARTS patterns for every fingerprint instead of keeping the patterns around for future use. This would probably improve the overall performance of FP…
-
https://wiki.lineageos.org/emulator
Would be nice to be able to download the compiled files directly like on Waydroid.
Tracking this issue on my Waydroid fork does not seem appropriate.
See h…
-
Any failure in SGMV comes back as `Request failed during generation: Server error: No suitable kernel. dtype=Half`
From Discord:
> I have tried the finetune adapter for llama2-7b. I trained mode…
-
### Steps to reproduce
1. Create a 1482x50 PNG image (can contain anything, width can be increased)
2. Send it in a room
3. Copy the event source contents, and remove the height information
4. S…
-
The quantization parameter `bnb_4bit_compute_dtype` seem to be optional for fine-tuning, but when we run lm-harness we get an error if it is not specified in the config.
Example: http://10.145.91.…
-
I use a OnePlus 5T with lineageOS (Android 14)
Inside the settings of any exercises, when I tap on what I suppose are drop-down menus (such as `Play cadance`, `Cadence Type`), the screen turns gray…