{ "cells": [ { "cell_type": "markdown", "id": "18662496-bb36-45a6-99ab-b2f0b91eb534", "metadata": {}, "source": [ "## Suno Demo\n", "\n", "Copy-Pasted from: https://colab.research.google.com/drive/1dWWkZzvu7L9Bunq9zvD-W02RFUXoW-Pd?usp=sharing#scrollTo=68QtoUqPWdLk\n" ] }, { "cell_type": "code", "execution_count": null, "id": "73cf9a1c-c6d4-492f-9d3d-8b19466e6014", "metadata": {}, "outputs": [], "source": [ "#!pip3 install optimum\n", "#!pip install -U flash-attn --no-build-isolation" ] }, { "cell_type": "code", "execution_count": 1, "id": "f6c0c08d-b1b7-479c-ae10-bd126d925bcd", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/home/sanyambhutani/.conda/envs/final-checking-meta/lib/python3.12/site-packages/transformers/models/encodec/modeling_encodec.py:124: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n", " self.register_buffer(\"padding_total\", torch.tensor(kernel_size - stride, dtype=torch.int64), persistent=False)\n" ] } ], "source": [ "from transformers import BarkModel, AutoProcessor\n", "import torch\n", "\n", "device = \"cuda:3\"\n", "\n", "processor = AutoProcessor.from_pretrained(\"suno/bark\")\n", "\n", "#model = model.to_bettertransformer()\n", "#model = BarkModel.from_pretrained(\"suno/bark\", torch_dtype=torch.float16, attn_implementation=\"flash_attention_2\").to(device)\n", "model = BarkModel.from_pretrained(\"suno/bark\", torch_dtype=torch.float16).to(device)#.to_bettertransformer()" ] }, { "cell_type": "code", "execution_count": 83, "id": "94c95582-49cf-419d-89bf-51e2e4e04379", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 83, "metadata": {}, "output_type": "execute_result" } ], "source": [ "voice_preset = \"v2/en_speaker_6\"\n", "\n", "# prepare the inputs\n", "text_prompt = \"\"\"\n", "[Laughs] Exactly! And the distillation part is where you take a large model and compress it down into a smaller, more efficient model that can run on devices with limited resources.\n", "\"\"\"\n", "inputs = processor(text_prompt, voice_preset=voice_preset)\n", "\n", "# generate speech\n", "speech_output = model.generate(**inputs.to(device))\n", "\n", "# let's hear it\n", "Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)" ] }, { "cell_type": "code", "execution_count": 9, "id": "a258e898-b007-4697-af9f-a9e7dbbb7fe4", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "speech_output = model.generate(**inputs, temperature = 0.7, semantic_temperature = 0.1)\n", "Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)" ] }, { "cell_type": "code", "execution_count": 10, "id": "11561846-d029-4f7e-be4a-4f7a09e84bf4", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "speech_output = model.generate(**inputs, temperature = 0.7, semantic_temperature = 0.2)\n", "Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)" ] }, { "cell_type": "code", "execution_count": 11, "id": "e80a13a5-6c5c-4850-bdc6-bf7d18484162", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "#I like this the most so far\n", "\n", "speech_output = model.generate(**inputs, temperature = 0.7, semantic_temperature = 0.3)\n", "Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)" ] }, { "cell_type": "code", "execution_count": 27, "id": "a8f3285e-d658-488f-a6e0-37f80c59cd04", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 27, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# This is better than 0.3\n", "\n", "speech_output = model.generate(**inputs, temperature = 0.7, semantic_temperature = 0.4)\n", "Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)" ] }, { "cell_type": "code", "execution_count": 13, "id": "b11d23af-62e3-432d-8ff5-0452c91cc42d", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 13, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Falls Apart\n", "\n", "speech_output = model.generate(**inputs, temperature = 0.7, semantic_temperature = 0.5)\n", "Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)" ] }, { "cell_type": "code", "execution_count": 28, "id": "045b9608-7ec9-4950-b3da-62d51ca3d792", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n", "Setting `pad_token_id` to `eos_token_id`:None for open-end generation.\n" ] }, { "data": { "text/html": [ "\n", "