-
### Summary
It has to be solved that deprecated TV types are not selectable anymore and not existing old TV types (i.e. removed custom TV types) could create a fatal error. Some fallback to a default…
-
**Describe the bug**
When running `LLAMA_CPP_LOG_LEVEL=4 cargo run --release --features rocm serve --model TabbyML/StarCoder-7B --device rocm`, the process hangs in the step where it tries to enable …
-
### :zap: Summary
The hardware platform suffix for Linux builds is now `amd64` or `arm64`. Change this for more reliable automation in provisioning.
### Value
Make container builds work w…
-
```
What steps will reproduce the problem?
1. Add the following to a gyp file to set multi-processor compilation.
'msvs_settings': {
'VCCLCompilerTool': {
'MultiProcessor…
-
```
What steps will reproduce the problem?
1. Add the following to a gyp file to set multi-processor compilation.
'msvs_settings': {
'VCCLCompilerTool': {
'MultiProcessor…
-
```
What steps will reproduce the problem?
1. Add the following to a gyp file to set multi-processor compilation.
'msvs_settings': {
'VCCLCompilerTool': {
'MultiProcessor…
-
```
What steps will reproduce the problem?
1. Add the following to a gyp file to set multi-processor compilation.
'msvs_settings': {
'VCCLCompilerTool': {
'MultiProcessor…
-
如题。看了看.md文档没能理解透彻,大大有配置好的方案可以分享吗?
-
Building on the amazing work by @mzbac and @nkasmanoff in https://github.com/ml-explore/mlx-examples/pull/461, I'd really love an example of how LLaVA 1.6 (aka llava next) can be fine-tuned with a LoR…
-
### Bug description
I'm trying to use the Streaming library by Mosaic as described in the doc (https://lightning.ai/docs/pytorch/stable/data/alternatives.html) but it doesn't work as seamlessly as ex…