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

Remove RWKV loader (#5130)

Browse source
This commit is contained in:
oobabooga 2023-12-31 02:01:40 -03:00 • committed by GitHub
parent 0e54a09bcb
commit 2734ce3e4c
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 3 additions and 189 deletions
Download patch file Download diff file Expand all files Collapse all files

5
modules/shared.py
View file

@ -165,11 +165,6 @@ group.add_argument('--deepspeed', action='store_true', help='Enable the use of D
group.add_argument('--nvme-offload-dir', type=str, help='DeepSpeed: Directory to use for ZeRO-3 NVME offloading.')
group.add_argument('--local_rank', type=int, default=0, help='DeepSpeed: Optional argument for distributed setups.')
# RWKV
group = parser.add_argument_group('RWKV')
group.add_argument('--rwkv-strategy', type=str, default=None, help='RWKV: The strategy to use while loading the model. Examples: "cpu fp32", "cuda fp16", "cuda fp16i8".')
group.add_argument('--rwkv-cuda-on', action='store_true', help='RWKV: Compile the CUDA kernel for better performance.')
# RoPE
group = parser.add_argument_group('RoPE')
group.add_argument('--alpha_value', type=float, default=1, help='Positional embeddings alpha factor for NTK RoPE scaling. Use either this or compress_pos_emb, not both.')