SOTA Open Source TTS
Fix document (use medium by default)
Lengyue9677932a21fd
diff --git a/docs/en/finetune.md b/docs/en/finetune.mdindex 694c39b..72db128 100644--- a/docs/en/finetune.md+++ b/docs/en/finetune.md@@ -154,7 +154,7 @@ huggingface-cli download fishaudio/fish-speech-1 text2semantic-sft-medium-v1-4k.Finally, you can start the fine-tuning by running the following command:```bashpython fish_speech/train.py --config-name text2semantic_finetune \- [email protected]=dual_ar_2_codebook_large+ [email protected]=dual_ar_2_codebook_medium```!!! note@@ -180,7 +180,7 @@ After training, you need to convert the LoRA weights to regular weights before p```bashpython tools/llama/merge_lora.py \- --llama-config dual_ar_2_codebook_large \+ --llama-config dual_ar_2_codebook_medium \--lora-config r_8_alpha_16 \--llama-weight checkpoints/text2semantic-sft-medium-v1-4k.pth \--lora-weight results/text2semantic-finetune-medium-lora/checkpoints/step_000000200.ckpt \diff --git a/docs/en/inference.md b/docs/en/inference.mdindex 4befc16..32973b0 100644--- a/docs/en/inference.md+++ b/docs/en/inference.md@@ -37,7 +37,7 @@ python tools/llama/generate.py \--text "The text you want to convert" \--prompt-text "Your reference text" \--prompt-tokens "fake.npy" \- --config-name dual_ar_2_codebook_large \+ --config-name dual_ar_2_codebook_medium \--checkpoint-path "checkpoints/text2semantic-sft-medium-v1-4k.pth" \--num-samples 2 \--compile@@ -70,7 +70,7 @@ We provide a HTTP API for inference. You can use the following command to startpython -m tools.api \--listen 0.0.0.0:8000 \--llama-checkpoint-path "checkpoints/text2semantic-sft-medium-v1-4k.pth" \- --llama-config-name dual_ar_2_codebook_large \+ --llama-config-name dual_ar_2_codebook_medium \--vqgan-checkpoint-path "checkpoints/vq-gan-group-fsq-2x1024.pth"```@@ -83,7 +83,7 @@ You can start the WebUI using the following command:```bashpython -m tools.webui \--llama-checkpoint-path "checkpoints/text2semantic-sft-medium-v1-4k.pth" \- --llama-config-name dual_ar_2_codebook_large \+ --llama-config-name dual_ar_2_codebook_medium \--vqgan-checkpoint-path "checkpoints/vq-gan-group-fsq-2x1024.pth"```diff --git a/docs/zh/finetune.md b/docs/zh/finetune.mdindex d4fde2f..213fa83 100644--- a/docs/zh/finetune.md+++ b/docs/zh/finetune.md@@ -165,7 +165,7 @@ HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech```bashpython fish_speech/train.py --config-name text2semantic_finetune \- [email protected]=dual_ar_2_codebook_large+ [email protected]=dual_ar_2_codebook_medium```!!! note@@ -190,7 +190,7 @@ python fish_speech/train.py --config-name text2semantic_finetune \```bashpython tools/llama/merge_lora.py \- --llama-config dual_ar_2_codebook_large \+ --llama-config dual_ar_2_codebook_medium \--lora-config r_8_alpha_16 \--llama-weight checkpoints/text2semantic-sft-medium-v1-4k.pth \--lora-weight results/text2semantic-finetune-medium-lora/checkpoints/step_000000200.ckpt \diff --git a/docs/zh/inference.md b/docs/zh/inference.mdindex 232088c..3b67142 100644--- a/docs/zh/inference.md+++ b/docs/zh/inference.md@@ -42,7 +42,7 @@ python tools/llama/generate.py \--text "要转换的文本" \--prompt-text "你的参考文本" \--prompt-tokens "fake.npy" \- --config-name dual_ar_2_codebook_large \+ --config-name dual_ar_2_codebook_medium \--checkpoint-path "checkpoints/text2semantic-sft-medium-v1-4k.pth" \--num-samples 2 \--compile@@ -75,7 +75,7 @@ python tools/vqgan/inference.py \python -m tools.api \--listen 0.0.0.0:8000 \--llama-checkpoint-path "checkpoints/text2semantic-sft-medium-v1-4k.pth" \- --llama-config-name dual_ar_2_codebook_large \+ --llama-config-name dual_ar_2_codebook_medium \--vqgan-checkpoint-path "checkpoints/vq-gan-group-fsq-2x1024.pth"# 推荐中国大陆用户运行以下命令来启动 HTTP 服务:@@ -91,7 +91,7 @@ HF_ENDPOINT=https://hf-mirror.com python -m ...```bashpython -m tools.webui \--llama-checkpoint-path "checkpoints/text2semantic-sft-medium-v1-4k.pth" \- --llama-config-name dual_ar_2_codebook_large \+ --llama-config-name dual_ar_2_codebook_medium \--vqgan-checkpoint-path "checkpoints/vq-gan-group-fsq-2x1024.pth"```diff --git a/tools/api.py b/tools/api.pyindex ff900b3..670c9bc 100644--- a/tools/api.py+++ b/tools/api.py@@ -227,7 +227,7 @@ def parse_args():default="checkpoints/text2semantic-sft-medium-v1-4k.pth",)parser.add_argument(- "--llama-config-name", type=str, default="dual_ar_2_codebook_large"+ "--llama-config-name", type=str, default="dual_ar_2_codebook_medium")parser.add_argument("--vqgan-checkpoint-path",diff --git a/tools/llama/merge_lora.py b/tools/llama/merge_lora.pyindex d125d12..5f9c2d0 100644--- a/tools/llama/merge_lora.py+++ b/tools/llama/merge_lora.py@@ -12,7 +12,7 @@ from fish_speech.models.text2semantic.lora_utils import (@click.command()[email protected]("--llama-config", type=str, default="dual_ar_2_codebook_large")[email protected]("--llama-config", type=str, default="dual_ar_2_codebook_medium")@click.option("--lora-config", type=str, default="r_8_alpha_16")@click.option("--llama-weight", type=str, default="checkpoints/text2semantic-sft-medium-v1-4k.pth"diff --git a/tools/webui.py b/tools/webui.pyindex 430fd1f..6186063 100644--- a/tools/webui.py+++ b/tools/webui.py@@ -324,7 +324,7 @@ def parse_args():default="checkpoints/text2semantic-sft-medium-v1-4k.pth",)parser.add_argument(- "--llama-config-name", type=str, default="dual_ar_2_codebook_large"+ "--llama-config-name", type=str, default="dual_ar_2_codebook_medium")parser.add_argument("--vqgan-checkpoint-path",