Add files using upload-large-folder tool

README.md

@@ -0,0 +1,110 @@
+---
+license: cc-by-nc-4.0
+library_name: woosh
+tags:
+  - audio
+  - audio-generation
+  - sound-effects
+  - foley
+  - text-to-audio
+  - video-to-audio
+  - t2a
+  - v2a
+  - safetensors
+  - woosh
+pipeline_tag: text-to-audio
+---
+
+# Woosh — Sony AI Sound-Effect Foundation Model (Mirror)
+
+This repository is a **community mirror** of the open weights released by Sony Research for [Woosh](https://github.com/SonyResearch/Woosh) — a foundation model for sound-effect generation supporting text-to-audio (T2A) and video-to-audio (V2A) synthesis.
+
+All files here are a one-to-one copy of Sony's [v1.0.0 GitHub release](https://github.com/SonyResearch/Woosh/releases/tag/v1.0.0), repackaged into a single browseable HF repo for convenience.
+
+## License — CC-BY-NC 4.0 (Non-Commercial)
+
+> All open weights in this repository are released by Sony Research under the [CC-BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) license. **Generated outputs inherit the non-commercial restriction.** You may not use model outputs in commercial products, paid releases, or client work. The upstream project's source code is released separately under MIT / Apache-2.0.
+
+If you need to attribute: **Sony Research — Woosh** ([arXiv / paper](https://github.com/SonyResearch/Woosh), GitHub: `SonyResearch/Woosh`).
+
+## Model Suite
+
+Woosh is a multi-model suite. All components are required together — the generative backbones depend on the shared AE, CLAP, and text conditioners at inference time.
+
+### Shared infrastructure
+
+| Folder | Role | File(s) |
+|---|---|---|
+| `checkpoints/Woosh-AE/` | Audio encoder / decoder producing high-quality latents | `weights.safetensors`, `config.yaml` |
+| `checkpoints/Woosh-CLAP/` | Multimodal text-audio alignment model (audio + text encoders) | `weights_audio.safetensors`, `weights_text.safetensors`, `config.yaml` |
+| `checkpoints/TextConditionerA/` | Text conditioner for the T2A path (pairs with Flow / DFlow) | `weights.safetensors`, `config.yaml` |
+| `checkpoints/TextConditionerV/` | Text conditioner for the V2A path (pairs with VFlow / DVFlow) | `weights.safetensors`, `config.yaml` |
+
+### Generative backbones
+
+| Folder | Task | Notes |
+|---|---|---|
+| `checkpoints/Woosh-Flow/` | Text → Audio | Full-quality T2A latent diffusion |
+| `checkpoints/Woosh-DFlow/` | Text → Audio | Distilled T2A — fewer steps, faster inference |
+| `checkpoints/Woosh-VFlow-8s/` | Video → Audio | V2A latent diffusion — **fixed 8-second output** |
+| `checkpoints/Woosh-DVFlow-8s/` | Video → Audio | Distilled V2A — fewer steps, **fixed 8-second output** |
+
+Every weight file ships as `safetensors`. No `.pt` / `.ckpt` / `.bin` in this mirror.
+
+## Layout
+
+```
+checkpoints/
+├── Woosh-AE/
+│   ├── weights.safetensors
+│   └── config.yaml
+├── Woosh-CLAP/
+│   ├── weights_audio.safetensors
+│   ├── weights_text.safetensors
+│   └── config.yaml
+├── TextConditionerA/
+│   ├── weights.safetensors
+│   └── config.yaml
+├── TextConditionerV/
+│   ├── weights.safetensors
+│   └── config.yaml
+├── Woosh-Flow/
+│   ├── weights.safetensors
+│   └── config.yaml
+├── Woosh-DFlow/
+│   ├── weights.safetensors
+│   └── config.yaml
+├── Woosh-VFlow-8s/
+│   ├── weights.safetensors
+│   └── config.yaml
+└── Woosh-DVFlow-8s/
+    ├── weights.safetensors
+    └── config.yaml
+```
+
+Directory names match Sony's release zip layout exactly so the upstream inference code finds its configs without modification.
+
+## Usage
+
+This mirror is intended to be consumed by [Sony's upstream `woosh` package](https://github.com/SonyResearch/Woosh). Clone and install the upstream repo, then point it at a local copy of this mirror's `checkpoints/` directory.
+
+```bash
+# Clone upstream
+git clone https://github.com/SonyResearch/Woosh.git
+cd Woosh
+
+# Sony's suggested env setup (uses uv)
+uv sync
+uv pip install -e .
+
+# Pull weights from this mirror
+hf download AEmotionStudio/woosh-models --local-dir ./
+```
+
+## Acknowledgements
+
+All credit for the Woosh models belongs to **Sony Research**. This mirror exists solely to make the CC-BY-NC open weights easier to fetch and integrate. Please cite the upstream project and respect the non-commercial license.
+
+- Upstream: <https://github.com/SonyResearch/Woosh>
+- Release: <https://github.com/SonyResearch/Woosh/releases/tag/v1.0.0>
+- License: [CC-BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/)

checkpoints/TextConditionerA/config.yaml

@@ -0,0 +1,22 @@
+exclude_from_checkpoint: false
+trainable: false
+sentence_config:
+  frozen: false
+  model: roberta-large
+  max_sentence_tokens: 77
+  adopt_n_layers: 0
+  adopt_layer_size: 2048
+  pool_type: eos
+  add_pooling_layer: true
+  hidden_dropout_prob: 0.1
+  attention_probs_dropout_prob: 0.1
+  finetune_n_layers: -1
+last_hidden_state: true
+use_shared_space: false
+normalize_shared_space: true
+freeze_clap: true
+lhs_index: -2
+remove_special_tokens: false
+eval_mode: true
+text_preprocessing: no_op
+shared_representation_size: 1024

checkpoints/TextConditionerA/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f66e1877efd49e3198068ffac74d00a84df18e53189643373e2e777224856bb
+size 1425689504

checkpoints/TextConditionerV/config.yaml

@@ -0,0 +1,22 @@
+exclude_from_checkpoint: false
+trainable: false
+sentence_config:
+  frozen: false
+  model: roberta-large
+  max_sentence_tokens: 77
+  adopt_n_layers: 0
+  adopt_layer_size: 2048
+  pool_type: eos
+  add_pooling_layer: true
+  hidden_dropout_prob: 0.1
+  attention_probs_dropout_prob: 0.1
+  finetune_n_layers: -1
+last_hidden_state: true
+use_shared_space: false
+normalize_shared_space: true
+freeze_clap: true
+lhs_index: -2
+remove_special_tokens: false
+eval_mode: true
+text_preprocessing: no_op
+shared_representation_size: 1024

checkpoints/TextConditionerV/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0dd912c322582cb95531de69cf1472c90752e468087a3b0fc461dea49975e80
+size 1425689504

checkpoints/Woosh-AE/config.yaml

@@ -0,0 +1,27 @@
+exclude_from_checkpoint: true
+trainable: true
+istft_head:
+  _target_: woosh.module.model.vocos.ISTFTCircleHead
+  _partial_: true
+  padding: center
+  conv_pad: same
+  conv_kernel: 1
+  softclip: softplus
+ztransform:
+  _target_: woosh.module.model.vocos.ZeroDropoutTransform
+  _partial_: true
+  p: 1.0
+_target_: woosh.module.model.VocosAutoEncoder
+channels: 1
+z_dim: 128
+d_model: 2048
+intermediate_dim: 3072
+n_fft: 960
+hop_length: 480
+num_layers: 8
+enc_num_layers: 8
+input_layer_norm: true
+final_layer_norm: true
+stft_normalized: false
+spec_embed: stft-complex
+sample_rate: 48000

checkpoints/Woosh-AE/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48e0d864beee90b17d232c74b16e9878dadc73b4ae55e4f6939df3f559c5d88a
+size 884664420

checkpoints/Woosh-CLAP/config.yaml

@@ -0,0 +1,25 @@
+shared_representation_size: 1024
+text_preprocessing: no_op
+audio:
+  name: passt_no
+  sample_rate: 32000
+  segment_length: 5
+  embedding_dim: 128
+  aggregate: mean
+  type: embedding
+  adopt_n_layers: 0
+  adopt_layer_size: 2048
+  s_patchout_t: 0
+  s_patchout_f: 0
+  eval_max_sec: 60
+  mask_attention: false
+sentence:
+  model: roberta-large
+  max_sentence_tokens: 77
+  adopt_n_layers: 0
+  adopt_layer_size: 2048
+  pool_type: eos
+  add_pooling_layer: true
+  hidden_dropout_prob: 0.1
+  attention_probs_dropout_prob: 0.1
+normalize: true

checkpoints/Woosh-CLAP/weights_audio.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d32f248f77f34cac2b6b71f13d4c1fb01207b1c83fc58dd130895c894f588d85
+size 347657136

checkpoints/Woosh-CLAP/weights_text.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f66e1877efd49e3198068ffac74d00a84df18e53189643373e2e777224856bb
+size 1425689504

checkpoints/Woosh-DFlow/config.yaml

@@ -0,0 +1,51 @@
+exclude_from_checkpoint: false
+trainable: true
+ldm:
+  exclude_from_checkpoint: false
+  trainable: true
+  model_type: LatentDiffusionModel
+  dit:
+    model_type: mmmssflux
+    rope_len_multiplier: 2
+    max_description_length: 77
+    max_seq_len: 501
+    dim: 1024
+    inter_dim: 4096
+    fixed_timestep_features: false
+    timestep_features_dim: 256
+    n_layers: 12
+    n_heads: 8
+    n_multimodal_layers: 6
+    qk_nope_head_dim: 16
+    qk_rope_head_dim: 112
+    qkv_head_dim: 128
+    n_memory_tokens_rope: 1
+    n_memory_tokens_description: 0
+    original_seq_len: 501
+    rope_theta: 10000.0
+    rope_factor: 40.0
+    beta_fast: 32
+    beta_slow: 1
+    io_channels: 128
+    cond_token_dim: 1024
+    adaln_last_layer: true
+    adaln_last_layer_nomod: false
+    non_checkpoint_layers: 14
+    mask_out_before: -1
+    estimate_logvar: true
+    no_description_mask: true
+    symmetric_attention_init: false
+    patch_size: 1
+    num_sinks: 0
+    mlp_act: gelu
+  conditioners:
+    text:
+      exclude_from_checkpoint: true
+      trainable: false
+      path: checkpoints/TextConditionerA
+  autoencoder:
+    exclude_from_checkpoint: true
+    trainable: false
+    path: checkpoints/Woosh-AE
+  sigma_data: 1.0
+  pred_type: v_pred

checkpoints/Woosh-DFlow/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fc4a2220fa4ba54e01611afe58e061854a58ad559eb56de2cca39112f7c8ef
+size 1378890620

checkpoints/Woosh-DVFlow-8s/config.yaml

@@ -0,0 +1,64 @@
+exclude_from_checkpoint: false
+trainable: true
+pretrained_model_type: videokontext
+ldm:
+  exclude_from_checkpoint: false
+  trainable: true
+  model_type: VideoKontextLDM
+  ldm:
+    exclude_from_checkpoint: false
+    trainable: true
+    dit:
+      model_type: mmmssflux
+      max_description_length: 77
+      max_seq_len: 501
+      rope_len_multiplier: 2
+      dim: 1024
+      inter_dim: 4096
+      fixed_timestep_features: false
+      timestep_features_dim: 256
+      n_layers: 12
+      n_heads: 8
+      n_multimodal_layers: 6
+      qk_nope_head_dim: 16
+      qk_rope_head_dim: 112
+      qkv_head_dim: 128
+      n_memory_tokens_rope: 1
+      n_memory_tokens_description: 0
+      original_seq_len: 501
+      rope_theta: 10000.0
+      rope_factor: 40.0
+      beta_fast: 32
+      beta_slow: 1
+      io_channels: 128
+      cond_token_dim: 1024
+      adaln_last_layer: true
+      adaln_last_layer_nomod: false
+      non_checkpoint_layers: 14
+      mask_out_before: -1
+      estimate_logvar: true
+      no_description_mask: true
+      symmetric_attention_init: false
+      patch_size: 1
+      num_sinks: 0
+      mlp_act: gelu
+    conditioners:
+      text:
+        exclude_from_checkpoint: true
+        trainable: false
+        path: checkpoints/TextConditionerV
+    autoencoder:
+      exclude_from_checkpoint: true
+      trainable: false
+      path: checkpoints/Woosh-AE
+    sigma_data: 1.0
+    pred_type: v_pred
+  audio_fps: 100
+  video_fps: 24
+  embed_dim: 768
+  embed_key: synch_out
+  pts_seconds_key: synch_pts_seconds
+  non_checkpoint_layers: 14
+  n_layers_encoder: 0
+  trainable_no_cond: false
+  use_batch_mask: true

checkpoints/Woosh-DVFlow-8s/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c87b796798fd96a4a3b1deeb70ccd05c6dc782fe546b622ce1368a05e441cd30
+size 1684261500

checkpoints/Woosh-Flow/config.yaml

@@ -0,0 +1,48 @@
+exclude_from_checkpoint: false
+trainable: true
+model_type: LatentDiffusionModel
+dit:
+  model_type: mmmssflux
+  rope_len_multiplier: 2
+  max_description_length: 77
+  max_seq_len: 501
+  dim: 1024
+  inter_dim: 4096
+  fixed_timestep_features: false
+  timestep_features_dim: 256
+  n_layers: 12
+  n_heads: 8
+  n_multimodal_layers: 6
+  qk_nope_head_dim: 16
+  qk_rope_head_dim: 112
+  qkv_head_dim: 128
+  n_memory_tokens_rope: 1
+  n_memory_tokens_description: 0
+  original_seq_len: 501
+  rope_theta: 10000.0
+  rope_factor: 40.0
+  beta_fast: 32
+  beta_slow: 1
+  io_channels: 128
+  cond_token_dim: 1024
+  adaln_last_layer: true
+  adaln_last_layer_nomod: false
+  non_checkpoint_layers: 14
+  mask_out_before: -1
+  estimate_logvar: true
+  no_description_mask: true
+  symmetric_attention_init: false
+  patch_size: 1
+  num_sinks: 0
+  mlp_act: gelu
+conditioners:
+  text:
+    exclude_from_checkpoint: true
+    trainable: false
+    path: checkpoints/TextConditionerA
+autoencoder:
+  exclude_from_checkpoint: true
+  trainable: false
+  path: checkpoints/Woosh-AE
+sigma_data: 1.0
+pred_type: v_pred

checkpoints/Woosh-Flow/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e2eafa2f9063934859888a2edd25ad4485f82dca4caaa7720f98d832f7ca39b
+size 1348979404

checkpoints/Woosh-VFlow-8s/config.yaml

@@ -0,0 +1,60 @@
+exclude_from_checkpoint: false
+trainable: true
+model_type: VideoKontextLDM
+ldm:
+  exclude_from_checkpoint: false
+  trainable: true
+  dit:
+    model_type: mmmssflux
+    max_description_length: 77
+    max_seq_len: 501
+    rope_len_multiplier: 2
+    dim: 1024
+    inter_dim: 4096
+    fixed_timestep_features: false
+    timestep_features_dim: 256
+    n_layers: 12
+    n_heads: 8
+    n_multimodal_layers: 6
+    qk_nope_head_dim: 16
+    qk_rope_head_dim: 112
+    qkv_head_dim: 128
+    n_memory_tokens_rope: 1
+    n_memory_tokens_description: 0
+    original_seq_len: 501
+    rope_theta: 10000.0
+    rope_factor: 40.0
+    beta_fast: 32
+    beta_slow: 1
+    io_channels: 128
+    cond_token_dim: 1024
+    adaln_last_layer: true
+    adaln_last_layer_nomod: false
+    non_checkpoint_layers: 14
+    mask_out_before: -1
+    estimate_logvar: true
+    no_description_mask: true
+    symmetric_attention_init: false
+    patch_size: 1
+    num_sinks: 0
+    mlp_act: gelu
+  conditioners:
+    text:
+      exclude_from_checkpoint: true
+      trainable: false
+      path: checkpoints/TextConditionerV
+  autoencoder:
+    exclude_from_checkpoint: true
+    trainable: false
+    path: checkpoints/Woosh-AE
+  sigma_data: 1.0
+  pred_type: v_pred
+audio_fps: 100
+video_fps: 24
+embed_dim: 768
+embed_key: synch_out
+pts_seconds_key: synch_pts_seconds
+non_checkpoint_layers: 14
+n_layers_encoder: 0
+trainable_no_cond: false
+use_batch_mask: true

checkpoints/Woosh-VFlow-8s/weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed2e7639e830060e532c955d4b4dc23869c8574605267846fa326c28730fcc7d
+size 1654350276