update checkpoint path

API_FLAGS.txt

@@ -1,6 +1,6 @@
 # --infer
 # --api
 --listen 0.0.0.0:8080 \
---llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
---decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+--llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+--decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
 --decoder-config-name firefly_gan_vq

docs/en/finetune.md

@@ -36,7 +36,7 @@ You need to convert your dataset into the above format and place it under `data`
 Make sure you have downloaded the VQGAN weights. If not, run the following command:
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 You can then run the following command to extract semantic tokens:
@@ -45,7 +45,7 @@ You can then run the following command to extract semantic tokens:
 python tools/vqgan/extract_vq.py data \
     --num-workers 1 --batch-size 16 \
     --config-name "firefly_gan_vq" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 !!! note
@@ -89,7 +89,7 @@ After the command finishes executing, you should see the `quantized-dataset-ft.p
 Similarly, make sure you have downloaded the `LLAMA` weights. If not, run the following command:
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 Finally, you can start the fine-tuning by running the following command:
@@ -117,9 +117,9 @@ After training, you need to convert the LoRA weights to regular weights before p
 ```bash
 python tools/llama/merge_lora.py \
 	--lora-config r_8_alpha_16 \
-	--base-weight checkpoints/fish-speech-1.2-sft \
+	--base-weight checkpoints/fish-speech-1.4 \
 	--lora-weight results/$project/checkpoints/step_000000010.ckpt \
-	--output checkpoints/fish-speech-1.2-sft-yth-lora/
+	--output checkpoints/fish-speech-1.4-yth-lora/
 ```
 !!! note
     You may also try other checkpoints. We suggest using the earliest checkpoint that meets your requirements, as they often perform better on out-of-distribution (OOD) data.

docs/en/inference.md

@@ -15,7 +15,7 @@ Inference support command line, HTTP API and web UI.
 Download the required `vqgan` and `llama` models from our Hugging Face repository.
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 ### 1. Generate prompt from voice:
@@ -26,7 +26,7 @@ huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/f
 ```bash
 python tools/vqgan/inference.py \
     -i "paimon.wav" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 You should get a `fake.npy` file.
@@ -38,7 +38,7 @@ python tools/llama/generate.py \
     --text "The text you want to convert" \
     --prompt-text "Your reference text" \
     --prompt-tokens "fake.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft" \
+    --checkpoint-path "checkpoints/fish-speech-1.4" \
     --num-samples 2 \
     --compile
 ```
@@ -59,7 +59,7 @@ This command will create a `codes_N` file in the working directory, where N is a
 ```bash
 python tools/vqgan/inference.py \
     -i "codes_0.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 ## HTTP API Inference
@@ -69,8 +69,8 @@ We provide a HTTP API for inference. You can use the following command to start
 ```bash
 python -m tools.api \
     --listen 0.0.0.0:8080 \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 
@@ -142,8 +142,8 @@ You can start the WebUI using the following command:
 
 ```bash
 python -m tools.webui \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 

docs/ja/finetune.md

@@ -36,7 +36,7 @@
 VQGANの重みをダウンロードしたことを確認してください。まだダウンロードしていない場合は、次のコマンドを実行してください。
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 次に、次のコマンドを実行してセマンティックトークンを抽出できます。
@@ -45,7 +45,7 @@ huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/f
 python tools/vqgan/extract_vq.py data \
     --num-workers 1 --batch-size 16 \
     --config-name "firefly_gan_vq" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 !!! note
@@ -89,7 +89,7 @@ python tools/llama/build_dataset.py \
 同様に、`LLAMA`の重みをダウンロードしたことを確認してください。まだダウンロードしていない場合は、次のコマンドを実行してください。
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 最後に、次のコマンドを実行して微調整を開始できます。
@@ -117,9 +117,9 @@ python fish_speech/train.py --config-name text2semantic_finetune \
 ```bash
 python tools/llama/merge_lora.py \
 	--lora-config r_8_alpha_16 \
-	--base-weight checkpoints/fish-speech-1.2-sft \
+	--base-weight checkpoints/fish-speech-1.4 \
 	--lora-weight results/$project/checkpoints/step_000000010.ckpt \
-	--output checkpoints/fish-speech-1.2-sft-yth-lora/
+	--output checkpoints/fish-speech-1.4-yth-lora/
 ```
 !!! note
     他のチェックポイントを試すこともできます。要件を満たす最も早いチェックポイントを使用することをお勧めします。これらは通常、分布外（OOD）データでより良いパフォーマンスを発揮します。

docs/ja/inference.md

@@ -15,7 +15,7 @@
 必要な`vqgan`および`llama`モデルを Hugging Face リポジトリからダウンロードします。
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 ### 1. 音声からプロンプトを生成する：
@@ -26,7 +26,7 @@ huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/f
 ```bash
 python tools/vqgan/inference.py \
     -i "paimon.wav" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 `fake.npy`ファイルが生成されるはずです。
@@ -38,7 +38,7 @@ python tools/llama/generate.py \
     --text "変換したいテキスト" \
     --prompt-text "参照テキスト" \
     --prompt-tokens "fake.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft" \
+    --checkpoint-path "checkpoints/fish-speech-1.4" \
     --num-samples 2 \
     --compile
 ```
@@ -59,7 +59,7 @@ python tools/llama/generate.py \
 ```bash
 python tools/vqgan/inference.py \
     -i "codes_0.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 ## HTTP API 推論
@@ -69,8 +69,8 @@ python tools/vqgan/inference.py \
 ```bash
 python -m tools.api \
     --listen 0.0.0.0:8080 \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 
@@ -146,8 +146,8 @@ python -m tools.post_api \
 
 ```bash
 python -m tools.webui \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 

docs/pt/finetune.md

@@ -36,7 +36,7 @@ Você precisa converter seu conjunto de dados para o formato acima e colocá-lo
 Certifique-se de ter baixado os pesos do VQGAN. Se não, execute o seguinte comando:
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 Em seguida, você pode executar o seguinte comando para extrair os tokens semânticos:
@@ -45,7 +45,7 @@ Em seguida, você pode executar o seguinte comando para extrair os tokens semân
 python tools/vqgan/extract_vq.py data \
     --num-workers 1 --batch-size 16 \
     --config-name "firefly_gan_vq" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 !!! note
@@ -89,7 +89,7 @@ Após executar o comando, você deverá ver o arquivo `quantized-dataset-ft.prot
 Da mesma forma, certifique-se de ter baixado os pesos do `LLAMA`. Se não, execute o seguinte comando:
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 E então, execute o seguinte comando para iniciar o ajuste fino:
@@ -117,9 +117,9 @@ Após o treinamento, é preciso converter os pesos do LoRA em pesos regulares an
 ```bash
 python tools/llama/merge_lora.py \
     --lora-config r_8_alpha_16 \
-    --base-weight checkpoints/fish-speech-1.2-sft \
+    --base-weight checkpoints/fish-speech-1.4 \
     --lora-weight results/$project/checkpoints/step_000000010.ckpt \
-    --output checkpoints/fish-speech-1.2-sft-yth-lora/
+    --output checkpoints/fish-speech-1.4-yth-lora/
 ```
 !!! note
     É possível também tentar outros checkpoints. Sugerimos usar o checkpoint que melhor atenda aos seus requisitos, pois eles geralmente têm um desempenho melhor em dados fora da distribuição (OOD).

docs/pt/inference.md

@@ -15,7 +15,7 @@ Suporte para inferência por linha de comando, API HTTP e interface web (WebUI).
 Baixe os modelos `vqgan` e `llama` necessários do nosso repositório Hugging Face.
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 ### 1. Gerar prompt a partir da voz:
@@ -26,7 +26,7 @@ huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/f
 ```bash
 python tools/vqgan/inference.py \
     -i "paimon.wav" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 Você deverá obter um arquivo `fake.npy`.
@@ -38,7 +38,7 @@ python tools/llama/generate.py \
     --text "O texto que você deseja converter" \
     --prompt-text "Seu texto de referência" \
     --prompt-tokens "fake.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft" \
+    --checkpoint-path "checkpoints/fish-speech-1.4" \
     --num-samples 2 \
     --compile
 ```
@@ -59,7 +59,7 @@ Este comando criará um arquivo `codes_N` no diretório de trabalho, onde N é u
 ```bash
 python tools/vqgan/inference.py \
     -i "codes_0.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 ## Inferência por API HTTP
@@ -69,8 +69,8 @@ Fornecemos uma API HTTP para inferência. O seguinte comando pode ser usado para
 ```bash
 python -m tools.api \
     --listen 0.0.0.0:8080 \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 
@@ -142,8 +142,8 @@ Para iniciar a WebUI de Inferência execute o seguinte comando:
 
 ```bash
 python -m tools.webui \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 

docs/zh/finetune.md

@@ -34,13 +34,13 @@
 确保你已经下载了 vqgan 权重, 如果没有, 请运行以下命令:
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 对于中国大陆用户, 可使用 mirror 下载.
 
 ```bash
-HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 随后可运行以下命令来提取语义 token:
@@ -49,7 +49,7 @@ HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech
 python tools/vqgan/extract_vq.py data \
     --num-workers 1 --batch-size 16 \
     --config-name "firefly_gan_vq" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 !!! note
@@ -93,13 +93,13 @@ python tools/llama/build_dataset.py \
 同样的, 请确保你已经下载了 `LLAMA` 权重, 如果没有, 请运行以下命令:
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 对于中国大陆用户, 可使用 mirror 下载.
 
 ```bash
-HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 最后, 你可以运行以下命令来启动微调:
@@ -127,9 +127,9 @@ python fish_speech/train.py --config-name text2semantic_finetune \
 ```bash
 python tools/llama/merge_lora.py \
 	--lora-config r_8_alpha_16 \
-	--base-weight checkpoints/fish-speech-1.2-sft \
+	--base-weight checkpoints/fish-speech-1.4 \
 	--lora-weight results/$project/checkpoints/step_000000010.ckpt \
-	--output checkpoints/fish-speech-1.2-sft-yth-lora/
+	--output checkpoints/fish-speech-1.4-yth-lora/
 ```
 
 !!! note

docs/zh/index.md

@@ -138,13 +138,13 @@ apt install libsox-dev
     确保您在 docker 容器内的终端，然后再从我们的 huggingface 仓库下载所需的 `vqgan` 和 `llama` 模型。
 
     ```bash
-    huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+    huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
     ```
 
     对于中国大陆用户，可以通过镜像站下载。
 
     ```bash
-    HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+    HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
     ```
 
 4. 配置环境变量，访问 WebUI

docs/zh/inference.md

@@ -15,13 +15,13 @@
 从我们的 huggingface 仓库下载所需的 `vqgan` 和 `llama` 模型。
 
 ```bash
-huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 对于中国大陆用户，可使用 mirror 下载。
 
 ```bash
-HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft
+HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4
 ```
 
 ### 1. 从语音生成 prompt:
@@ -32,7 +32,7 @@ HF_ENDPOINT=https://hf-mirror.com huggingface-cli download fishaudio/fish-speech
 ```bash
 python tools/vqgan/inference.py \
     -i "paimon.wav" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 你应该能得到一个 `fake.npy` 文件.
@@ -44,7 +44,7 @@ python tools/llama/generate.py \
     --text "要转换的文本" \
     --prompt-text "你的参考文本" \
     --prompt-tokens "fake.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft" \
+    --checkpoint-path "checkpoints/fish-speech-1.4" \
     --num-samples 2 \
     --compile
 ```
@@ -65,7 +65,7 @@ python tools/llama/generate.py \
 ```bash
 python tools/vqgan/inference.py \
     -i "codes_0.npy" \
-    --checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    --checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
 ```
 
 ## HTTP API 推理
@@ -75,8 +75,8 @@ python tools/vqgan/inference.py \
 ```bash
 python -m tools.api \
     --listen 0.0.0.0:8080 \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 如果你想要加速推理，可以加上`--compile`参数。
@@ -153,8 +153,8 @@ python -m tools.post_api \
 
 ```bash
 python -m tools.webui \
-    --llama-checkpoint-path "checkpoints/fish-speech-1.2-sft" \
-    --decoder-checkpoint-path "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth" \
+    --llama-checkpoint-path "checkpoints/fish-speech-1.4" \
+    --decoder-checkpoint-path "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth" \
     --decoder-config-name firefly_gan_vq
 ```
 

fish_speech/configs/text2semantic_finetune.yaml

@@ -4,7 +4,7 @@ defaults:
 
 project: text2semantic_finetune_dual_ar
 max_length: 4096
-pretrained_ckpt_path: checkpoints/fish-speech-1.2-sft
+pretrained_ckpt_path: checkpoints/fish-speech-1.4
 
 # Lightning Trainer
 trainer:

fish_speech/webui/manage.py

@@ -469,7 +469,7 @@ def train_process(
                 "--config-name",
                 "firefly_gan_vq",
                 "--checkpoint-path",
-                "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+                "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
             ]
         )
 
@@ -485,7 +485,7 @@ def train_process(
                 "16",
             ]
         )
-        ckpt_path = "checkpoints/fish-speech-1.2-sft/model.pth"
+        ckpt_path = "checkpoints/fish-speech-1.4/model.pth"
         lora_prefix = "lora_" if llama_use_lora else ""
         llama_name = lora_prefix + "text2semantic_" + new_project
         latest = next(
@@ -925,9 +925,9 @@ with gr.Blocks(
                                         "Type the path or select from the dropdown"
                                     ),
                                     choices=[
-                                        "checkpoints/fish-speech-1.2-sft/model.pth",
+                                        "checkpoints/fish-speech-1.4/model.pth",
                                     ],
-                                    value="checkpoints/fish-speech-1.2-sft/model.pth",
+                                    value="checkpoints/fish-speech-1.4/model.pth",
                                     allow_custom_value=True,
                                     interactive=True,
                                 )
@@ -979,7 +979,7 @@ with gr.Blocks(
                                         "Type the path or select from the dropdown"
                                     ),
                                     choices=list_llama_models(),
-                                    value="checkpoints/fish-speech-1.2-sft",
+                                    value="checkpoints/fish-speech-1.4",
                                     allow_custom_value=True,
                                     interactive=True,
                                 )
@@ -1042,7 +1042,7 @@ with gr.Blocks(
                                         "Type the path or select from the dropdown"
                                     ),
                                     choices=list_decoder_models(),
-                                    value="checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+                                    value="checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
                                     allow_custom_value=True,
                                 )
                                 infer_decoder_config = gr.Dropdown(
@@ -1060,7 +1060,7 @@ with gr.Blocks(
                                     info=i18n(
                                         "Type the path or select from the dropdown"
                                     ),
-                                    value="checkpoints/fish-speech-1.2-sft",
+                                    value="checkpoints/fish-speech-1.4",
                                     choices=list_llama_models(),
                                     allow_custom_value=True,
                                 )

inference.ipynb

@@ -61,7 +61,7 @@
     "# !set HF_ENDPOINT=https://hf-mirror.com\n",
     "# !export HF_ENDPOINT=https://hf-mirror.com \n",
     "\n",
-    "!huggingface-cli download fishaudio/fish-speech-1.2-sft --local-dir checkpoints/fish-speech-1.2-sft/"
+    "!huggingface-cli download fishaudio/fish-speech-1.4 --local-dir checkpoints/fish-speech-1.4/"
    ]
   },
   {
@@ -84,8 +84,8 @@
    "outputs": [],
    "source": [
     "!python tools/webui.py \\\n",
-    "    --llama-checkpoint-path checkpoints/fish-speech-1.2-sft \\\n",
-    "    --decoder-checkpoint-path checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth \\\n",
+    "    --llama-checkpoint-path checkpoints/fish-speech-1.4 \\\n",
+    "    --decoder-checkpoint-path checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth \\\n",
     "    # --compile"
    ]
   },
@@ -122,7 +122,7 @@
     "\n",
     "!python tools/vqgan/inference.py \\\n",
     "    -i {src_audio} \\\n",
-    "    --checkpoint-path \"checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth\"\n",
+    "    --checkpoint-path \"checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth\"\n",
     "\n",
     "from IPython.display import Audio, display\n",
     "audio = Audio(filename=\"fake.wav\")\n",
@@ -158,7 +158,7 @@
     "    --text \"hello world\" \\\n",
     "    --prompt-text \"The text corresponding to reference audio\" \\\n",
     "    --prompt-tokens \"fake.npy\" \\\n",
-    "    --checkpoint-path \"checkpoints/fish-speech-1.2-sft\" \\\n",
+    "    --checkpoint-path \"checkpoints/fish-speech-1.4\" \\\n",
     "    --num-samples 2\n",
     "    # --compile"
    ]
@@ -182,7 +182,7 @@
    "source": [
     "!python tools/vqgan/inference.py \\\n",
     "    -i \"codes_0.npy\" \\\n",
-    "    --checkpoint-path \"checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth\"\n",
+    "    --checkpoint-path \"checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth\"\n",
     "\n",
     "from IPython.display import Audio, display\n",
     "audio = Audio(filename=\"fake.wav\")\n",

pyproject.toml

@@ -56,4 +56,4 @@ requires = ["setuptools", "setuptools-scm"]
 build-backend = "setuptools.build_meta"
 
 [tool.setuptools]
-packages = ["fish_speech"]
+packages = ["fish_speech", "tools"]

tools/api.py

@@ -345,12 +345,12 @@ def parse_args():
     parser.add_argument(
         "--llama-checkpoint-path",
         type=str,
-        default="checkpoints/fish-speech-1.2-sft",
+        default="checkpoints/fish-speech-1.4",
     )
     parser.add_argument(
         "--decoder-checkpoint-path",
         type=str,
-        default="checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+        default="checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
     )
     parser.add_argument("--decoder-config-name", type=str, default="firefly_gan_vq")
     parser.add_argument("--device", type=str, default="cuda")

tools/download_models.py

@@ -22,8 +22,8 @@ def check_and_download_files(repo_id, file_list, local_dir):
 
 
 # 1st
-repo_id_1 = "fishaudio/fish-speech-1.2-sft"
-local_dir_1 = "./checkpoints/fish-speech-1.2-sft"
+repo_id_1 = "fishaudio/fish-speech-1.4"
+local_dir_1 = "./checkpoints/fish-speech-1.4"
 files_1 = [
     "model.pth",
     "README.md",
@@ -31,7 +31,7 @@ files_1 = [
     "tokenizer_config.json",
     "tokenizer.json",
     "config.json",
-    "firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+    "firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
 ]
 
 # 3rd

tools/llama/generate.py

@@ -612,7 +612,7 @@ def launch_thread_safe_queue(
 @click.option(
     "--checkpoint-path",
     type=click.Path(path_type=Path, exists=True),
-    default="checkpoints/fish-speech-1.2-sft",
+    default="checkpoints/fish-speech-1.4",
 )
 @click.option("--device", type=str, default="cuda")
 @click.option("--compile/--no-compile", default=False)

tools/llama/merge_lora.py

@@ -15,7 +15,7 @@ from fish_speech.models.text2semantic.lora import get_merged_state_dict
 
 @click.command()
 @click.option("--lora-config", type=str, default="r_8_alpha_16")
-@click.option("--base-weight", type=str, default="checkpoints/fish-speech-1.2-sft")
+@click.option("--base-weight", type=str, default="checkpoints/fish-speech-1.4")
 @click.option("--lora-weight", type=str, required=True)
 @click.option("--output", type=str, required=True)
 def merge(lora_config, base_weight, lora_weight, output):

tools/llama/quantize.py

@@ -428,7 +428,7 @@ def generate_folder_name():
 @click.option(
     "--checkpoint-path",
     type=click.Path(path_type=Path, exists=True),
-    default="checkpoints/fish-speech-1.2-sft",
+    default="checkpoints/fish-speech-1.4",
 )
 @click.option(
     "--mode", type=str, default="int8", help="type of quantization to perform"
@@ -451,7 +451,7 @@ def quantize(checkpoint_path: Path, mode: str, groupsize: int, timestamp: str) -
         precision=precision,
         compile=False,
     )
-    vq_model = "firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
+    vq_model = "firefly-gan-vq-fsq-8x1024-21hz-generator.pth"
     now = timestamp if timestamp != "None" else generate_folder_name()
 
     if mode == "int8":

tools/vqgan/extract_vq.py

@@ -42,7 +42,7 @@ logger.add(sys.stderr, format=logger_format)
 @lru_cache(maxsize=1)
 def get_model(
     config_name: str = "firefly_gan_vq",
-    checkpoint_path: str = "checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+    checkpoint_path: str = "checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
     device: str | torch.device = "cuda",
 ):
     with initialize(version_base="1.3", config_path="../../fish_speech/configs"):
@@ -133,7 +133,7 @@ def process_batch(files: list[Path], model) -> float:
 @click.option("--config-name", default="firefly_gan_vq")
 @click.option(
     "--checkpoint-path",
-    default="checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+    default="checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
 )
 @click.option("--batch-size", default=64)
 @click.option("--filelist", default=None, type=Path)

tools/vqgan/inference.py

@@ -59,7 +59,7 @@ def load_model(config_name, checkpoint_path, device="cuda"):
 @click.option("--config-name", default="firefly_gan_vq")
 @click.option(
     "--checkpoint-path",
-    default="checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+    default="checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
 )
 @click.option(
     "--device",

tools/webui.py

@@ -560,12 +560,12 @@ def parse_args():
     parser.add_argument(
         "--llama-checkpoint-path",
         type=Path,
-        default="checkpoints/fish-speech-1.2-sft",
+        default="checkpoints/fish-speech-1.4",
     )
     parser.add_argument(
         "--decoder-checkpoint-path",
         type=Path,
-        default="checkpoints/fish-speech-1.2-sft/firefly-gan-vq-fsq-4x1024-42hz-generator.pth",
+        default="checkpoints/fish-speech-1.4/firefly-gan-vq-fsq-8x1024-21hz-generator.pth",
     )
     parser.add_argument("--decoder-config-name", type=str, default="firefly_gan_vq")
     parser.add_argument("--device", type=str, default="cuda")