Replicate Update: Support Multiple Images

[asutermo]

Hi there, I was keen on using replicate and generating multiple images in one go. I've made this PR to allow the usage of multiple images.

• Add param 'num_images_per_prompt' to schema
• Return a list of Paths for each image
• Rename images to out_{iteration}.png

Follow up:

Given more time, I'd probably look at updating OmnigenPipeline to do this too.

Sample Output Locally:

I cleaned the following output a touch (removed base64 output strings)

curl http://localhost:5003/predictions -X POST --header "Content-Type: application/json" --data '{"input": {"prompt": "Create a scenic view of the mountains", "num_images_per_prompt": 3, "width": 128, "height": 128}}'
{"input":{"prompt":"Create a scenic view of the mountains","img1":null,"img2":null,"img3":null,"width":128,"height":128,"inference_steps":50,"guidance_scale":2.5,"img_guidance_scale":1.6,"seed":null,"max_input_image_size":1024,"separate_cfg_infer":true,"offload_model":false,"use_input_image_size_as_output":false,"num_images_per_prompt":3},"output":["data:image/png;base64,iV...=="],"id":null,"version":null,"created_at":null,"started_at":"2025-01-10T21:07:46.098716+00:00","completed_at":"2025-01-10T21:08:05.749434+00:00","logs":"Using seed: 1497\n  0%|          | 0/50 [00:00<?, ?it/s]\n  2%|▏         | 1/50 [00:00<00:32,  1.50it/s]\n  6%|▌         | 3/50 [00:00<00:10,  4.31it/s]\n 10%|█         | 5/50 [00:01<00:07,  6.11it/s]\n 12%|█▏        | 6/50 [00:01<00:06,  6.81it/s]\n 16%|█▌        | 8/50 [00:01<00:04,  8.64it/s]\n 20%|██        | 10/50 [00:01<00:04,  9.77it/s]\n 24%|██▍       | 12/50 [00:01<00:03, 10.39it/s]\n 28%|██▊       | 14/50 [00:02<00:04,  7.53it/s]\n 30%|███       | 15/50 [00:02<00:05,  6.92it/s]\n 32%|███▏      | 16/50 [00:02<00:05,  6.41it/s]\n 34%|███▍      | 17/50 [00:02<00:05,  5.95it/s]\n 36%|███▌      | 18/50 [00:02<00:04,  6.50it/s]\n 40%|████      | 20/50 [00:02<00:03,  8.13it/s]\n 44%|████▍     | 22/50 [00:03<00:02,  9.52it/s]\n 48%|████▊     | 24/50 [00:03<00:02, 10.42it/s]\n 52%|█████▏    | 26/50 [00:03<00:02, 11.16it/s]\n 56%|█████▌    | 28/50 [00:03<00:01, 11.77it/s]\n 60%|██████    | 30/50 [00:03<00:01, 12.04it/s]\n 64%|██████▍   | 32/50 [00:03<00:01, 12.54it/s]\n 68%|██████▊   | 34/50 [00:04<00:01, 10.11it/s]\n 72%|███████▏  | 36/50 [00:04<00:01, 10.20it/s]\n 76%|███████▌  | 38/50 [00:04<00:01, 10.90it/s]\n 80%|████████  | 40/50 [00:04<00:00, 11.49it/s]\n 84%|████████▍ | 42/50 [00:04<00:00, 11.99it/s]\n 88%|████████▊ | 44/50 [00:04<00:00, 12.21it/s]\n 92%|█████████▏| 46/50 [00:05<00:00, 12.34it/s]\n 96%|█████████▌| 48/50 [00:05<00:00, 12.51it/s]\n100%|██████████| 50/50 [00:05<00:00, 12.76it/s]\n100%|██████████| 50/50 [00:05<00:00,  9.34it/s]\n  0%|
 | 0/50 [00:00<?, ?it/s]\n  2%|▏         | 1/50 [00:00<00:07,  6.95it/s]\n  6%|▌         | 3/50 [00:00<00:04, 10.14it/s]\n 10%|█         | 5/50 [00:00<00:03, 11.41it/s]\n 14%|█▍        | 7/50 [00:00<00:03, 11.34it/s]\n 18%|█▊        | 9/50 [00:00<00:03, 11.87it/s]\n 22%|██▏       | 11/50 [00:00<00:03, 12.11it/s]\n 26%|██▌       | 13/50 [00:01<00:03, 12.32it/s]\n 30%|███       | 15/50 [00:01<00:02, 12.72it/s]\n 34%|███▍      | 17/50 [00:01<00:02, 12.86it/s]\n 38%|███▊      | 19/50 [00:01<00:02, 12.87it/s]\n 42%|████▏     | 21/50 [00:01<00:02, 13.13it/s]\n 46%|████▌     | 23/50 [00:01<00:02, 13.13it/s]\n 50%|█████     | 25/50 [00:02<00:01, 12.95it/s]\n 54%|█████▍    | 27/50 [00:02<00:01, 12.87it/s]\n 58%|█████▊    | 29/50 [00:02<00:01, 12.62it/s]\n 62%|██████▏   | 31/50 [00:02<00:01, 12.76it/s]\n 66%|██████▌   | 33/50 [00:02<00:01,  8.84it/s]\n 70%|███████   | 35/50 [00:03<00:02,  7.34it/s]\n 74%|███████▍  | 37/50 [00:03<00:01,  8.56it/s]\n 78%|███████▊  | 39/50 [00:03<00:01,  9.39it/s]\n 82%|████████▏ | 41/50 [00:03<00:00,  9.08it/s]\n 86%|████████▌ | 43/50 [00:04<00:00,  8.77it/s]\n 90%|█████████ | 45/50 [00:04<00:00,  9.61it/s]\n 94%|█████████▍| 47/50 [00:04<00:00, 10.49it/s]\n 98%|█████████▊| 49/50 [00:04<00:00,  8.50it/s]\n100%|██████████| 50/50 [00:04<00:00, 10.44it/s]\n  0%|          | 0/50 [00:00<?, ?it/s]\n  2%|▏         | 1/50 [00:00<00:06,  7.40it/s]\n  6%|▌         | 3/50 [00:00<00:04,  9.76it/s]\n 10%|█         | 5/50 [00:00<00:04, 10.69it/s]\n 14%|█▍        | 7/50 [00:00<00:03, 11.71it/s]\n 18%|█▊        | 9/50 [00:00<00:03, 12.17it/s]\n 22%|██▏       | 11/50 [00:00<00:03, 11.13it/s]\n 26%|██▌       | 13/50 [00:01<00:03, 11.90it/s]\n 30%|███       | 15/50 [00:01<00:03, 10.83it/s]\n 34%|███▍      | 17/50 [00:01<00:03, 10.54it/s]\n 38%|███▊      | 19/50 [00:01<00:02, 10.86it/s]\n 42%|████▏     | 21/50 [00:01<00:02, 11.55it/s]\n 46%|████▌     | 23/50 [00:02<00:02, 12.01it/s]\n 50%|█████     | 25/50 [00:02<00:02, 12.43it/s]\n 54%|█████▍    | 27/50 [00:02<00:01, 12.20it/s]\n 58%|█████▊    | 29/50 [00:02<00:01, 11.79it/s]\n 62%|██████▏   | 31/50 [00:02<00:01, 11.05it/s]\n 66%|██████▌   | 33/50 [00:02<00:01, 10.22it/s]\n 70%|███████   | 35/50 [00:03<00:01,  9.97it/s]\n 74%|███████▍  | 37/50 [00:03<00:01, 10.55it/s]\n 78%|███████▊  | 39/50 [00:03<00:00, 11.35it/s]\n 82%|████████▏ | 41/50 [00:03<00:00, 11.83it/s]\n 86%|████████▌ | 43/50 [00:03<00:00, 11.37it/s]\n 90%|█████████ | 45/50 [00:03<00:00, 11.78it/s]\n 94%|█████████▍| 47/50 [00:04<00:00, 12.15it/s]\n 98%|█████████▊| 49/50 [00:04<00:00, 12.43it/s]\n100%|██████████| 50/50 [00:04<00:00, 11.43it/s]\n","error":null,"status":"succeeded","metrics":{"predict_time":19.650718}}

Using seed: 1497
  0%|          | 0/50 [00:00<?, ?it/s]
  2%|▏         | 1/50 [00:00<00:32,  1.50it/s]
  6%|▌         | 3/50 [00:00<00:10,  4.31it/s]
 10%|█         | 5/50 [00:01<00:07,  6.11it/s]
 12%|█▏        | 6/50 [00:01<00:06,  6.81it/s]
 16%|█▌        | 8/50 [00:01<00:04,  8.64it/s]
 20%|██        | 10/50 [00:01<00:04,  9.77it/s]
 24%|██▍       | 12/50 [00:01<00:03, 10.39it/s]
 28%|██▊       | 14/50 [00:02<00:04,  7.53it/s]
 30%|███       | 15/50 [00:02<00:05,  6.92it/s]
 32%|███▏      | 16/50 [00:02<00:05,  6.41it/s]
 34%|███▍      | 17/50 [00:02<00:05,  5.95it/s]
 36%|███▌      | 18/50 [00:02<00:04,  6.50it/s]
 40%|████      | 20/50 [00:02<00:03,  8.13it/s]
 44%|████▍     | 22/50 [00:03<00:02,  9.52it/s]
 48%|████▊     | 24/50 [00:03<00:02, 10.42it/s]
 52%|█████▏    | 26/50 [00:03<00:02, 11.16it/s]
 56%|█████▌    | 28/50 [00:03<00:01, 11.77it/s]
 60%|██████    | 30/50 [00:03<00:01, 12.04it/s]
 64%|██████▍   | 32/50 [00:03<00:01, 12.54it/s]
 68%|██████▊   | 34/50 [00:04<00:01, 10.11it/s]
 72%|███████▏  | 36/50 [00:04<00:01, 10.20it/s]
 76%|███████▌  | 38/50 [00:04<00:01, 10.90it/s]
 80%|████████  | 40/50 [00:04<00:00, 11.49it/s]
 84%|████████▍ | 42/50 [00:04<00:00, 11.99it/s]
 88%|████████▊ | 44/50 [00:04<00:00, 12.21it/s]
 92%|█████████▏| 46/50 [00:05<00:00, 12.34it/s]
 96%|█████████▌| 48/50 [00:05<00:00, 12.51it/s]
100%|██████████| 50/50 [00:05<00:00, 12.76it/s]
100%|██████████| 50/50 [00:05<00:00,  9.34it/s]
  0%|          | 0/50 [00:00<?, ?it/s]
  2%|▏         | 1/50 [00:00<00:07,  6.95it/s]
  6%|▌         | 3/50 [00:00<00:04, 10.14it/s]
 10%|█         | 5/50 [00:00<00:03, 11.41it/s]
 14%|█▍        | 7/50 [00:00<00:03, 11.34it/s]
 18%|█▊        | 9/50 [00:00<00:03, 11.87it/s]
 22%|██▏       | 11/50 [00:00<00:03, 12.11it/s]
 26%|██▌       | 13/50 [00:01<00:03, 12.32it/s]
 30%|███       | 15/50 [00:01<00:02, 12.72it/s]
 34%|███▍      | 17/50 [00:01<00:02, 12.86it/s]
 38%|███▊      | 19/50 [00:01<00:02, 12.87it/s]
 42%|████▏     | 21/50 [00:01<00:02, 13.13it/s]
 46%|████▌     | 23/50 [00:01<00:02, 13.13it/s]
 50%|█████     | 25/50 [00:02<00:01, 12.95it/s]
 54%|█████▍    | 27/50 [00:02<00:01, 12.87it/s]
 58%|█████▊    | 29/50 [00:02<00:01, 12.62it/s]
 62%|██████▏   | 31/50 [00:02<00:01, 12.76it/s]
 66%|██████▌   | 33/50 [00:02<00:01,  8.84it/s]
 70%|███████   | 35/50 [00:03<00:02,  7.34it/s]
 74%|███████▍  | 37/50 [00:03<00:01,  8.56it/s]
 78%|███████▊  | 39/50 [00:03<00:01,  9.39it/s]
 82%|████████▏ | 41/50 [00:03<00:00,  9.08it/s]
 86%|████████▌ | 43/50 [00:04<00:00,  8.77it/s]
 90%|█████████ | 45/50 [00:04<00:00,  9.61it/s]
 94%|█████████▍| 47/50 [00:04<00:00, 10.49it/s]
 98%|█████████▊| 49/50 [00:04<00:00,  8.50it/s]
100%|██████████| 50/50 [00:04<00:00, 10.44it/s]
  0%|          | 0/50 [00:00<?, ?it/s]
  2%|▏         | 1/50 [00:00<00:06,  7.40it/s]
  6%|▌         | 3/50 [00:00<00:04,  9.76it/s]
 10%|█         | 5/50 [00:00<00:04, 10.69it/s]
 14%|█▍        | 7/50 [00:00<00:03, 11.71it/s]
 18%|█▊        | 9/50 [00:00<00:03, 12.17it/s]
 22%|██▏       | 11/50 [00:00<00:03, 11.13it/s]
 26%|██▌       | 13/50 [00:01<00:03, 11.90it/s]
 30%|███       | 15/50 [00:01<00:03, 10.83it/s]
 34%|███▍      | 17/50 [00:01<00:03, 10.54it/s]
 38%|███▊      | 19/50 [00:01<00:02, 10.86it/s]
 42%|████▏     | 21/50 [00:01<00:02, 11.55it/s]
 46%|████▌     | 23/50 [00:02<00:02, 12.01it/s]
 50%|█████     | 25/50 [00:02<00:02, 12.43it/s]
 54%|█████▍    | 27/50 [00:02<00:01, 12.20it/s]
 58%|█████▊    | 29/50 [00:02<00:01, 11.79it/s]
 62%|██████▏   | 31/50 [00:02<00:01, 11.05it/s]
 66%|██████▌   | 33/50 [00:02<00:01, 10.22it/s]
 70%|███████   | 35/50 [00:03<00:01,  9.97it/s]
 74%|███████▍  | 37/50 [00:03<00:01, 10.55it/s]
 78%|███████▊  | 39/50 [00:03<00:00, 11.35it/s]
 82%|████████▏ | 41/50 [00:03<00:00, 11.83it/s]
 86%|████████▌ | 43/50 [00:03<00:00, 11.37it/s]
 90%|█████████ | 45/50 [00:03<00:00, 11.78it/s]
 94%|█████████▍| 47/50 [00:04<00:00, 12.15it/s]
 98%|█████████▊| 49/50 [00:04<00:00, 12.43it/s]
100%|██████████| 50/50 [00:04<00:00, 11.43it/s]
{"prediction_id": null, "logger": "cog.server.runner", "timestamp": "2025-01-10T21:08:05.749089Z", "severity": "INFO", "message": "prediction succeeded"}