The Fish Speech V1.4 model. (Updated 4 months, 3 weeks ago)
This is a modal window.
Beginning of dialog window. Escape will cancel and close the window.
End of dialog window.
{
"text": "With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.",
"text_reference": "and keeping eternity before the eyes, though much",
"speaker_reference": "https://replicate.delivery/pbxt/MNFXdPaUPOwYCZjZM4azsymbzE2TCV2WJXfGpeV2DrFWaSq8/example_en.wav"
}npm install replicateimport Replicate from "replicate";
import fs from "node:fs";
const replicate = new Replicate({
auth: process.env.REPLICATE_API_TOKEN,
});Run ttsds/fishspeech_1_4 using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.
const output = await replicate.run(
"ttsds/fishspeech_1_4:7d55af8314c9ec4206d76c1e958cd8807c9c1bd59bffcfec363aea89e7179dd8",
{
input: {
text: "With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.",
text_reference: "and keeping eternity before the eyes, though much",
speaker_reference: "https://replicate.delivery/pbxt/MNFXdPaUPOwYCZjZM4azsymbzE2TCV2WJXfGpeV2DrFWaSq8/example_en.wav"
}
}
);
// To access the file URL:
console.log(output.url()); //=> "http://example.com"
// To write the file to disk:
fs.writeFile("my-image.png", output);To learn more, take a look at the guide on getting started with Node.js.
pip install replicateimport replicateRun ttsds/fishspeech_1_4 using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.
output = replicate.run(
"ttsds/fishspeech_1_4:7d55af8314c9ec4206d76c1e958cd8807c9c1bd59bffcfec363aea89e7179dd8",
input={
"text": "With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.",
"text_reference": "and keeping eternity before the eyes, though much",
"speaker_reference": "https://replicate.delivery/pbxt/MNFXdPaUPOwYCZjZM4azsymbzE2TCV2WJXfGpeV2DrFWaSq8/example_en.wav"
}
)
print(output)To learn more, take a look at the guide on getting started with Python.
Run ttsds/fishspeech_1_4 using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.
curl -s -X POST \
-H "Authorization: Bearer $REPLICATE_API_TOKEN" \
-H "Content-Type: application/json" \
-H "Prefer: wait" \
-d $'{
"version": "ttsds/fishspeech_1_4:7d55af8314c9ec4206d76c1e958cd8807c9c1bd59bffcfec363aea89e7179dd8",
"input": {
"text": "With tenure, Suzie\'d have all the more leisure for yachting, but her publications are no good.",
"text_reference": "and keeping eternity before the eyes, though much",
"speaker_reference": "https://replicate.delivery/pbxt/MNFXdPaUPOwYCZjZM4azsymbzE2TCV2WJXfGpeV2DrFWaSq8/example_en.wav"
}
}' \
https://api.replicate.com/v1/predictionsTo learn more, take a look at Replicate’s HTTP API reference docs.
This is a modal window.
Beginning of dialog window. Escape will cancel and close the window.
End of dialog window.
{
"completed_at": "2025-01-28T18:47:31.189703Z",
"created_at": "2025-01-28T18:46:12.997000Z",
"data_removed": false,
"error": null,
"id": "ct6fzs250nrma0cmnrybghpfxg",
"input": {
"text": "With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.",
"text_reference": "and keeping eternity before the eyes, though much",
"speaker_reference": "https://replicate.delivery/pbxt/MNFXdPaUPOwYCZjZM4azsymbzE2TCV2WJXfGpeV2DrFWaSq8/example_en.wav"
},
"logs": "2025-01-28 18:47:27.414 | INFO | tools.llama.generate:generate_long:759 - Encoded text: With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.\n2025-01-28 18:47:27.415 | INFO | tools.llama.generate:generate_long:777 - Generating sentence 1/1 of sample 1/1\n 0%| | 0/3965 [00:00<?, ?it/s]/root/.pyenv/versions/3.11.10/lib/python3.11/site-packages/torch/backends/cuda/__init__.py:342: FutureWarning: torch.backends.cuda.sdp_kernel() is deprecated. In the future, this context manager will be removed. Please see, torch.nn.attention.sdpa_kernel() for the new context manager, with updated signature.\nwarnings.warn(\n 0%| | 4/3965 [00:00<01:44, 37.72it/s]\n 0%| | 8/3965 [00:00<01:44, 38.01it/s]\n 0%| | 12/3965 [00:00<01:43, 38.31it/s]\n 0%| | 16/3965 [00:00<01:46, 37.10it/s]\n 1%| | 20/3965 [00:00<01:49, 35.87it/s]\n 1%| | 24/3965 [00:00<01:47, 36.77it/s]\n 1%| | 28/3965 [00:00<01:45, 37.39it/s]\n 1%| | 32/3965 [00:00<01:43, 37.86it/s]\n 1%| | 36/3965 [00:00<01:43, 38.08it/s]\n 1%| | 40/3965 [00:01<01:42, 38.33it/s]\n 1%| | 44/3965 [00:01<01:42, 38.35it/s]\n 1%| | 48/3965 [00:01<01:41, 38.49it/s]\n 1%|▏ | 52/3965 [00:01<01:41, 38.62it/s]\n 1%|▏ | 56/3965 [00:01<01:40, 38.71it/s]\n 2%|▏ | 60/3965 [00:01<01:40, 38.74it/s]\n 2%|▏ | 64/3965 [00:01<01:40, 38.75it/s]\n 2%|▏ | 68/3965 [00:01<01:40, 38.76it/s]\n 2%|▏ | 72/3965 [00:01<01:40, 38.74it/s]\n 2%|▏ | 76/3965 [00:01<01:40, 38.72it/s]\n 2%|▏ | 80/3965 [00:02<01:41, 38.29it/s]\n 2%|▏ | 84/3965 [00:02<01:41, 38.16it/s]\n 2%|▏ | 88/3965 [00:02<01:41, 38.27it/s]\n 2%|▏ | 92/3965 [00:02<01:40, 38.45it/s]\n 2%|▏ | 96/3965 [00:02<01:41, 38.12it/s]\n 3%|▎ | 100/3965 [00:02<01:40, 38.28it/s]\n 3%|▎ | 104/3965 [00:02<01:40, 38.39it/s]\n 3%|▎ | 108/3965 [00:02<01:40, 38.53it/s]\n 3%|▎ | 112/3965 [00:02<01:39, 38.65it/s]\n 3%|▎ | 116/3965 [00:03<01:39, 38.70it/s]\n 3%|▎ | 120/3965 [00:03<01:39, 38.66it/s]\n3%|▎ | 122/3965 [00:03<01:41, 37.95it/s]\n2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:832 - Generated 124 tokens in 3.41 seconds, 36.34 tokens/sec\n2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:835 - Bandwidth achieved: 17.97 GB/s\n2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:840 - GPU Memory used: 1.63 GB\nNext sample",
"metrics": {
"predict_time": 4.110862369,
"total_time": 78.192703
},
"output": "https://replicate.delivery/xezq/Owfq9xLg25zBPqSGVQTaT7Drf28prPfFwe4RjEq1XvLPE2mQB/generated.wav",
"started_at": "2025-01-28T18:47:27.078840Z",
"status": "succeeded",
"urls": {
"stream": "https://stream.replicate.com/v1/files/bsvm-7g4dwjmh2xs55r4h4pcijvspuc7vgens5p4jgbuggjgwea5uawfa",
"get": "https://api.replicate.com/v1/predictions/ct6fzs250nrma0cmnrybghpfxg",
"cancel": "https://api.replicate.com/v1/predictions/ct6fzs250nrma0cmnrybghpfxg/cancel"
},
"version": "7d55af8314c9ec4206d76c1e958cd8807c9c1bd59bffcfec363aea89e7179dd8"
}2025-01-28 18:47:27.414 | INFO | tools.llama.generate:generate_long:759 - Encoded text: With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.
2025-01-28 18:47:27.415 | INFO | tools.llama.generate:generate_long:777 - Generating sentence 1/1 of sample 1/1
0%| | 0/3965 [00:00<?, ?it/s]/root/.pyenv/versions/3.11.10/lib/python3.11/site-packages/torch/backends/cuda/__init__.py:342: FutureWarning: torch.backends.cuda.sdp_kernel() is deprecated. In the future, this context manager will be removed. Please see, torch.nn.attention.sdpa_kernel() for the new context manager, with updated signature.
warnings.warn(
0%| | 4/3965 [00:00<01:44, 37.72it/s]
0%| | 8/3965 [00:00<01:44, 38.01it/s]
0%| | 12/3965 [00:00<01:43, 38.31it/s]
0%| | 16/3965 [00:00<01:46, 37.10it/s]
1%| | 20/3965 [00:00<01:49, 35.87it/s]
1%| | 24/3965 [00:00<01:47, 36.77it/s]
1%| | 28/3965 [00:00<01:45, 37.39it/s]
1%| | 32/3965 [00:00<01:43, 37.86it/s]
1%| | 36/3965 [00:00<01:43, 38.08it/s]
1%| | 40/3965 [00:01<01:42, 38.33it/s]
1%| | 44/3965 [00:01<01:42, 38.35it/s]
1%| | 48/3965 [00:01<01:41, 38.49it/s]
1%|▏ | 52/3965 [00:01<01:41, 38.62it/s]
1%|▏ | 56/3965 [00:01<01:40, 38.71it/s]
2%|▏ | 60/3965 [00:01<01:40, 38.74it/s]
2%|▏ | 64/3965 [00:01<01:40, 38.75it/s]
2%|▏ | 68/3965 [00:01<01:40, 38.76it/s]
2%|▏ | 72/3965 [00:01<01:40, 38.74it/s]
2%|▏ | 76/3965 [00:01<01:40, 38.72it/s]
2%|▏ | 80/3965 [00:02<01:41, 38.29it/s]
2%|▏ | 84/3965 [00:02<01:41, 38.16it/s]
2%|▏ | 88/3965 [00:02<01:41, 38.27it/s]
2%|▏ | 92/3965 [00:02<01:40, 38.45it/s]
2%|▏ | 96/3965 [00:02<01:41, 38.12it/s]
3%|▎ | 100/3965 [00:02<01:40, 38.28it/s]
3%|▎ | 104/3965 [00:02<01:40, 38.39it/s]
3%|▎ | 108/3965 [00:02<01:40, 38.53it/s]
3%|▎ | 112/3965 [00:02<01:39, 38.65it/s]
3%|▎ | 116/3965 [00:03<01:39, 38.70it/s]
3%|▎ | 120/3965 [00:03<01:39, 38.66it/s]
3%|▎ | 122/3965 [00:03<01:41, 37.95it/s]
2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:832 - Generated 124 tokens in 3.41 seconds, 36.34 tokens/sec
2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:835 - Bandwidth achieved: 17.97 GB/s
2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:840 - GPU Memory used: 1.63 GB
Next sampleWant to make some of these yourself?
Run this modelThis model is cold. You'll get a fast response if the model is warm and already running, and a slower response if the model is cold and starting up.
This model runs on A100 (80GB). View more.
2025-01-28 18:47:27.414 | INFO | tools.llama.generate:generate_long:759 - Encoded text: With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.
2025-01-28 18:47:27.415 | INFO | tools.llama.generate:generate_long:777 - Generating sentence 1/1 of sample 1/1
0%| | 0/3965 [00:00<?, ?it/s]/root/.pyenv/versions/3.11.10/lib/python3.11/site-packages/torch/backends/cuda/__init__.py:342: FutureWarning: torch.backends.cuda.sdp_kernel() is deprecated. In the future, this context manager will be removed. Please see, torch.nn.attention.sdpa_kernel() for the new context manager, with updated signature.
warnings.warn(
0%| | 4/3965 [00:00<01:44, 37.72it/s]
0%| | 8/3965 [00:00<01:44, 38.01it/s]
0%| | 12/3965 [00:00<01:43, 38.31it/s]
0%| | 16/3965 [00:00<01:46, 37.10it/s]
1%| | 20/3965 [00:00<01:49, 35.87it/s]
1%| | 24/3965 [00:00<01:47, 36.77it/s]
1%| | 28/3965 [00:00<01:45, 37.39it/s]
1%| | 32/3965 [00:00<01:43, 37.86it/s]
1%| | 36/3965 [00:00<01:43, 38.08it/s]
1%| | 40/3965 [00:01<01:42, 38.33it/s]
1%| | 44/3965 [00:01<01:42, 38.35it/s]
1%| | 48/3965 [00:01<01:41, 38.49it/s]
1%|▏ | 52/3965 [00:01<01:41, 38.62it/s]
1%|▏ | 56/3965 [00:01<01:40, 38.71it/s]
2%|▏ | 60/3965 [00:01<01:40, 38.74it/s]
2%|▏ | 64/3965 [00:01<01:40, 38.75it/s]
2%|▏ | 68/3965 [00:01<01:40, 38.76it/s]
2%|▏ | 72/3965 [00:01<01:40, 38.74it/s]
2%|▏ | 76/3965 [00:01<01:40, 38.72it/s]
2%|▏ | 80/3965 [00:02<01:41, 38.29it/s]
2%|▏ | 84/3965 [00:02<01:41, 38.16it/s]
2%|▏ | 88/3965 [00:02<01:41, 38.27it/s]
2%|▏ | 92/3965 [00:02<01:40, 38.45it/s]
2%|▏ | 96/3965 [00:02<01:41, 38.12it/s]
3%|▎ | 100/3965 [00:02<01:40, 38.28it/s]
3%|▎ | 104/3965 [00:02<01:40, 38.39it/s]
3%|▎ | 108/3965 [00:02<01:40, 38.53it/s]
3%|▎ | 112/3965 [00:02<01:39, 38.65it/s]
3%|▎ | 116/3965 [00:03<01:39, 38.70it/s]
3%|▎ | 120/3965 [00:03<01:39, 38.66it/s]
3%|▎ | 122/3965 [00:03<01:41, 37.95it/s]
2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:832 - Generated 124 tokens in 3.41 seconds, 36.34 tokens/sec
2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:835 - Bandwidth achieved: 17.97 GB/s
2025-01-28 18:47:30.827 | INFO | tools.llama.generate:generate_long:840 - GPU Memory used: 1.63 GB
Next sample