Duplicate from deCoold/flowAI_API

Co-authored-by: Andrea De Cosmo <deCoold@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,37 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+*.pt.tmp filter=lfs diff=lfs merge=lfs -text
+*.tmp filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -0,0 +1,23 @@
+# Use the official Python image
+FROM python:3.9
+
+# Set the working directory
+WORKDIR /app
+
+# Copy the requirements file
+COPY requirements.txt requirements.txt
+
+# Copy the application code
+COPY app.py app.py
+
+# Clone the shap-e repository
+RUN git clone https://github.com/openai/shap-e.git
+
+# Install dependencies
+RUN pip install -r requirements.txt
+
+# Expose the desired port (same as in your app.py)
+EXPOSE 5000
+
+# Set the entry point command
+CMD python ./app.py

README.md

@@ -0,0 +1,12 @@
+---
+title: FlowAI API
+emoji: 👁
+colorFrom: green
+colorTo: green
+sdk: docker
+pinned: false
+license: apache-2.0
+duplicated_from: deCoold/flowAI_API
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,87 @@
+from flask import Flask, request, send_file
+from flask_cors import CORS
+
+
+import torch
+
+
+import shap_e
+from shap_e.diffusion.sample import sample_latents
+from shap_e.diffusion.gaussian_diffusion import diffusion_from_config
+from shap_e.models.download import load_model, load_config
+from shap_e.util.notebooks import create_pan_cameras, decode_latent_images, decode_latent_mesh
+
+
+import os
+
+
+# To remove on diploy
+#os.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'max_split_size_mb:768'
+
+app = Flask(__name__)
+
+app.config['CORS_HEADERS'] = 'Content-Type'
+cors = CORS(app, resorces={r'/generate_3d': {"origins": '*'}})
+
+
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+
+print('Device loaded: ', device)
+print('Loading models...')
+
+path_file = os.path.abspath(os.path.join(os.path.dirname(__file__), 'shape_e_model_cache'))
+trasmitter_path = os.path.join(path_file, 'transmitter')
+
+xm = load_model(trasmitter_path, device=device)
+model = load_model('text300M', device=device)
+diffusion = diffusion_from_config(load_config('diffusion'))
+
+@app.route('/generate_3d', methods=['POST', 'OPTIONS'])
+def generate_3d():
+
+    if request.method == 'OPTIONS':
+        return '', 200
+
+    print('Generating 3D...')
+
+    batch_size = 1
+    guidance_scale = 15.0 
+    prompt = request.json['prompt'] 
+
+    latents = sample_latents(
+        batch_size=batch_size,
+        model=model,
+        diffusion=diffusion,
+        guidance_scale=guidance_scale,
+        model_kwargs=dict(texts=[prompt] * batch_size),
+        progress=True,
+        clip_denoised=True,
+        use_fp16=True,
+        use_karras=True,
+        karras_steps=64,
+        sigma_min=1E-3,
+        sigma_max=160,
+        s_churn=0,
+    )
+
+    render_mode = 'nerf'  
+    size = 64
+
+    cameras = create_pan_cameras(size, device)
+    for i, latent in enumerate(latents):
+        images = decode_latent_images(xm, latent, cameras, rendering_mode=render_mode)
+
+    filename = f'tmp_mesh.obj'
+    
+    t = decode_latent_mesh(xm, latents[0]).tri_mesh()
+    with open(filename, 'w') as f: 
+        t.write_obj(f)
+
+    print('3D asset generated')
+
+    return send_file(filename, as_attachment=True)
+
+
+if __name__ == '__main__':
+    app.run(port=5000, debug=True)

requirements.txt

@@ -0,0 +1,7 @@
+flask==2.1.0
+PyYAML
+ipywidgets==7.6.5
+flask-cors
+torch
+
+-e ./shap-e

shap-e

@@ -0,0 +1 @@
+Subproject commit 0b26c720b28fe31f8a79fc5a9dc907fbd0f67162

shap_e_model_cache/ViT-L-14.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8cca3fd41ae0c99ba7e8951adf17d267cdb84cd88be6f7c2e0eca1737a03836
+size 932768134

shap_e_model_cache/diffusion_config.yaml

@@ -0,0 +1,3 @@
+mean_type: x_start
+schedule: exp
+timesteps: 1024

shap_e_model_cache/text_cond.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6b4fa599a7b3c3b16c222d5f5fe56f9db9289ff0b6575fbe5c11bc97106aad4
+size 1262868003

shap_e_model_cache/text_cond_config.yaml

@@ -0,0 +1,14 @@
+d_latent: 1048576
+inner:
+  cond_drop_prob: 0.1
+  heads: 16
+  init_scale: 0.25
+  layers: 24
+  name: CLIPImagePointDiffusionTransformer
+  pos_emb_init_scale: 0.05
+  time_token_cond: true
+  token_cond: true
+  use_pos_emb: true
+  width: 1024
+latent_ctx: 1024
+name: SplitVectorDiffusion

shap_e_model_cache/transmitter.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af02a0b85a8abdfb3919584b63c540ba175f6ad4790f574a7fef4617e5acdc3b
+size 1776048343

shap_e_model_cache/transmitter.pt.tmp

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb2cccc4de381a23364782390ab87e67168d00511ae8a12031540edbf40735c6
+size 106119168

shap_e_model_cache/transmitter_config.yaml

@@ -0,0 +1,80 @@
+encoder:
+  cross_attention_dataset: pcl_and_multiview_pcl
+  d_latent: 1024
+  data_ctx: 1024
+  fps_method: first
+  heads: 8
+  init_scale: 0.25
+  inner_batch_size:
+  - 16384
+  - 20
+  input_channels: 6
+  latent_bottleneck:
+    diffusion:
+      schedule: inv_parabola
+      schedule_args:
+        power: 5.0
+      timesteps: 1024
+    diffusion_prob: 0.1
+    name: clamp_diffusion_noise
+  layers: 12
+  max_depth: 9.0
+  max_unrolls: 1
+  min_unrolls: 1
+  name: PointCloudPerceiverChannelsEncoder
+  params_proj:
+    init_scale: 1.0
+    learned_scale: 0.0625
+    name: channels
+    use_ln: true
+  patch_size: 8
+  pointconv_hidden:
+  - 1024
+  - 1024
+  pointconv_padding_mode: circular
+  pointconv_patch_size: 8
+  pointconv_samples: 64
+  pointconv_stride: 4
+  pos_emb: nerf
+  use_depth: true
+  use_pointconv: true
+  width: 1024
+name: Transmitter
+renderer:
+  grid_size: 128
+  n_coarse_samples: 64
+  n_fine_samples: 128
+  name: NeRSTFRenderer
+  nerstf:
+    activation: swish
+    d_hidden: 256
+    density_activation: relu
+    init_scale: 0.25
+    initial_density_bias: 0.1
+    insert_direction_at: 4
+    meta_bias: false
+    meta_parameters: true
+    n_hidden_layers: 6
+    n_meta_layers: 4
+    name: MLPNeRSTFModel
+    posenc_version: nerf
+    separate_coarse_channels: true
+    separate_nerf_channels: true
+    trainable_meta: false
+  separate_shared_samples: true
+  void:
+    background:
+    - 0
+    - 0
+    - 0
+    name: VoidNeRFModel
+  volume:
+    bbox_max:
+    - 1.0
+    - 1.0
+    - 1.0
+    bbox_min:
+    - -1.0
+    - -1.0
+    - -1.0
+    name: BoundingBoxVolume

to_deploy.md

@@ -0,0 +1,3 @@
+git add Dockerfile
+git commit -m "Add application file"
+git push