Text-to-Speech
Transformers
Safetensors
parallel_s2s
text-generation
speech
spoken-language-model
audio
tts
asr
flexislm
Instructions to use FlexiSLM/FlexiSLM-7B-Stage2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use FlexiSLM/FlexiSLM-7B-Stage2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-to-speech", model="FlexiSLM/FlexiSLM-7B-Stage2")# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("FlexiSLM/FlexiSLM-7B-Stage2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "AUD_END_TOKEN": 151938, | |
| "AUD_START_TOKEN": 151937, | |
| "AUD_TAG_TOKEN": 151939, | |
| "adaptor_input_dim": 5120, | |
| "add_length_embeddings": true, | |
| "alignment_text_pad_token_id": 151643, | |
| "architectures": [ | |
| "ParallelS2SForCausalLM" | |
| ], | |
| "assistant_text_start_delay_tokens": -1, | |
| "attention_dropout": 0.0, | |
| "audio_embed_mlp_dropout": 0.0, | |
| "audio_embed_mlp_hidden_ratio": 4.0, | |
| "audio_end_id": null, | |
| "audio_vocab_size": 32768, | |
| "chained_3b_model_path": null, | |
| "chained_adaptor_hidden_size": 512, | |
| "codec_hidden_size": 3584, | |
| "dtype": "bfloat16", | |
| "early_diverge_talker": false, | |
| "enable_flexible_framerate": true, | |
| "eos_token_id": 151645, | |
| "finetune_speech_encoder": false, | |
| "flexicodec_ckpt_path": "/F00120260003/flexislm_project/model/FlexiCodec/nartts_flexicodec_only.safetensors", | |
| "flexicodec_config_path": "/F00120260003/flexislm_project/model/FlexiCodec/12hz_v1_half_config.yaml", | |
| "flow_matching_decoder_ckpt_path": "/F00120260003/flexislm_project/model/FlexiCodec/nartts.safetensors", | |
| "flow_matching_vocoder_path": "/F00120260003/flexislm_project/model/FlexiCodec/vocos_emilia.safetensors", | |
| "force_use_combined_embedding": false, | |
| "framerate_max": 1.0, | |
| "framerate_min": 0.0, | |
| "framerate_token_id": null, | |
| "freeze_llm": false, | |
| "freeze_talker": false, | |
| "hidden_act": "silu", | |
| "hidden_size": 3584, | |
| "initializer_range": 0.02, | |
| "input_merging_transformer_causal": false, | |
| "input_merging_transformer_context": 32, | |
| "input_merging_transformer_d_model": 768, | |
| "input_merging_transformer_dim_feedforward": 2048, | |
| "input_merging_transformer_num_heads": 8, | |
| "input_merging_transformer_num_layers": 4, | |
| "intermediate_size": 18944, | |
| "layer_types": [ | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention" | |
| ], | |
| "length_loss_weight": 0.1, | |
| "lora_alpha": 64, | |
| "lora_rank": 32, | |
| "max_length_classes": 32, | |
| "max_position_embeddings": 32768, | |
| "max_tokens_per_group": 16, | |
| "max_window_layers": 28, | |
| "model_type": "parallel_s2s", | |
| "no_pad": true, | |
| "num_attention_heads": 28, | |
| "num_hidden_layers": 28, | |
| "num_key_value_heads": 4, | |
| "num_prefix_tokens": 32, | |
| "only_train_llm": false, | |
| "only_train_talker": false, | |
| "output_uniform_merging": false, | |
| "pad_token_id": 151643, | |
| "padded_audio_vocab_size": 32768, | |
| "per_sample_frame_rate_embed": true, | |
| "predict_second_audio_token": false, | |
| "qwen25o_encoder_config_path": "/F00120260003/flexislm_project/model/Qwen2.5-Omni-7B/config.json", | |
| "qwen25o_encoder_path": "/F00120260003/flexislm_project/model/Qwen2.5-Omni-7B", | |
| "qwen3_encoder_config_path": null, | |
| "qwen3_encoder_path": null, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 1000000.0, | |
| "sensevoice_small_path": "/F00120260003/flexislm_project/model/SenseVoiceSmall", | |
| "sliding_window": null, | |
| "speech_delay_tokens": 5, | |
| "talker_concat_lm_text_output": true, | |
| "talker_embed_v2": false, | |
| "talker_hidden_size": 1280, | |
| "talker_intermediate_size": 5120, | |
| "talker_num_attention_heads": 8, | |
| "talker_num_layers": 20, | |
| "talker_pretrained_model_path": null, | |
| "text_alignment_pad_loss_weight": null, | |
| "text_loss_weight": 2.0, | |
| "text_vocab_size": 151940, | |
| "thinker_concat_user_speech": false, | |
| "tie_word_embeddings": false, | |
| "training_framerate_options": [ | |
| 0.89, | |
| 0.91, | |
| 1.0 | |
| ], | |
| "training_input_framerate_options": [ | |
| 5.0, | |
| 5.5, | |
| 6.0, | |
| 6.5, | |
| 7.0, | |
| 7.5, | |
| 8.0, | |
| 8.5, | |
| 9.0, | |
| 9.5, | |
| 10.0, | |
| 11.0, | |
| 12.0, | |
| 12.5 | |
| ], | |
| "transformers_version": "4.57.3", | |
| "uniform_merging": false, | |
| "use_attention_gating": false, | |
| "use_cache": true, | |
| "use_chained_architecture": false, | |
| "use_combined_embedding": false, | |
| "use_concat_len_emb": false, | |
| "use_input_merging_transformer": true, | |
| "use_input_merging_transformer_v2": true, | |
| "use_joint_text_audio_vocab": false, | |
| "use_learnable_audio_boundary": true, | |
| "use_learnable_prefix": false, | |
| "use_lora": true, | |
| "use_mlp_for_audio_embed": true, | |
| "use_omni_token": true, | |
| "use_qwen25o_feature": true, | |
| "use_qwen3_feature": false, | |
| "use_sensevoice_feature": false, | |
| "use_sinusoidal": true, | |
| "use_sliding_window": false, | |
| "use_whisper_feature": false, | |
| "use_whisper_fetaure": false, | |
| "vocab_size": 151944, | |
| "whisper_encoder_path": null | |
| } | |