mirror of https://github.com/ollama/ollama.git
* use ggml_*_split activations when possible * forward qkv |
||
|---|---|---|
| .. | ||
| ggml | ||
| ggml.go | ||
| quantization.go | ||
| threads.go | ||
| threads_debug.go | ||
* use ggml_*_split activations when possible * forward qkv |
||
|---|---|---|
| .. | ||
| ggml | ||
| ggml.go | ||
| quantization.go | ||
| threads.go | ||
| threads_debug.go | ||