mmarch/text-generation-webui
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Add the --cpu option for llama.cpp to prevent CUDA from being used (#3432)

Browse source
This commit is contained in:
oobabooga 2023-08-03 11:00:36 -03:00 • committed by GitHub
parent 3e70bce576
commit 87dab03dc0
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 40 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

5
README.md
View file

@ -249,8 +249,9 @@ Optionally, you can use the following command-line flags:
| `--n-gpu-layers N_GPU_LAYERS` | Number of layers to offload to the GPU. Only works if llama-cpp-python was compiled with BLAS. Set this to 1000000000 to offload all layers to the GPU. |
| `--n_ctx N_CTX` | Size of the prompt context. |
| `--llama_cpp_seed SEED` | Seed for llama-cpp models. Default 0 (random). |
| `--n_gqa N_GQA` | grouped-query attention. Must be 8 for llama2 70b. |
| `--rms_norm_eps RMS_NORM_EPS` | Must be 1e-5 for llama2 70b. |
| `--n_gqa N_GQA` | grouped-query attention. Must be 8 for llama-2 70b. |
| `--rms_norm_eps RMS_NORM_EPS` | 5e-6 is a good value for llama-2 models. |
| `--cpu` | Use the CPU version of llama-cpp-python instead of the GPU-accelerated version. |
#### AutoGPTQ