diff --git a/README.md b/README.md index 13268352..822deba9 100644 --- a/README.md +++ b/README.md @@ -337,9 +337,9 @@ Optionally, you can use the following command-line flags: | Flag | Description | |------------------|-------------| -| `--alpha_value ALPHA_VALUE` | Positional embeddings alpha factor for NTK RoPE scaling. Use either this or compress_pos_emb, not both. -| `--rope_freq_base ROPE_FREQ_BASE` | If greater than 0, will be used instead of alpha_value. Those two are related by rope_freq_base = 10000 * alpha_value ^ (64 / 63) -| `--compress_pos_emb COMPRESS_POS_EMB` | Positional embeddings compression factor. Should be set to (context length) / (model's original context length). Equal to 1/rope_freq_scale. +| `--alpha_value ALPHA_VALUE` | Positional embeddings alpha factor for NTK RoPE scaling. Use either this or compress_pos_emb, not both. | +| `--rope_freq_base ROPE_FREQ_BASE` | If greater than 0, will be used instead of alpha_value. Those two are related by rope_freq_base = 10000 * alpha_value ^ (64 / 63). | +| `--compress_pos_emb COMPRESS_POS_EMB` | Positional embeddings compression factor. Should be set to (context length) / (model's original context length). Equal to 1/rope_freq_scale. | #### Gradio diff --git a/modules/shared.py b/modules/shared.py index 49557a8d..ca68c1a6 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -160,7 +160,7 @@ parser.add_argument('--rwkv-cuda-on', action='store_true', help='RWKV: Compile t # RoPE parser.add_argument('--alpha_value', type=int, default=1, help="Positional embeddings alpha factor for NTK RoPE scaling. Use either this or compress_pos_emb, not both.") -parser.add_argument('--rope_freq_base', type=int, default=1, help="If greater than 0, will be used instead of alpha_value. Those two are related by rope_freq_base = 10000 * alpha_value ^ (64 / 63)") +parser.add_argument('--rope_freq_base', type=int, default=0, help="If greater than 0, will be used instead of alpha_value. Those two are related by rope_freq_base = 10000 * alpha_value ^ (64 / 63).") parser.add_argument('--compress_pos_emb', type=int, default=1, help="Positional embeddings compression factor. Should be set to (context length) / (model\'s original context length). Equal to 1/rope_freq_scale.") # Gradio