5 Model added

templates/CLIP-ViT-Large/README.md

@@ -0,0 +1,19 @@
+# CLIP ViT-Large/14
+
+OpenAI's powerful vision-language model for understanding images and text in a shared embedding space.
+
+Unleash the power of zero-shot image classification with Nosana! Run CLIP on GPU-backed nodes for instant image recognition without any task-specific training.
+
+## Key Features
+- Vision Transformer (ViT) architecture with 428M parameters
+- Zero-shot image classification and retrieval
+- Joint image-text embedding space
+- State-of-the-art performance on diverse benchmarks
+- Robust to distribution shifts
+
+## Configuration
+- Port: 9000
+- GPU: Required (4GB VRAM)
+- REST API for image classification and embedding
+- Supports both vision and text modalities
+- Handles arbitrary image classification tasks

templates/CLIP-ViT-Large/info.json

@@ -0,0 +1,9 @@
+{
+  "id": "clip-vit-large",
+  "name": "CLIP ViT-Large/14",
+  "description": "OpenAI's Vision-Language model for zero-shot image classification and embedding",
+  "category": ["Image classification", "New", "Multimodal"],
+  "icon": "https://storage.googleapis.com/sf-blog-images/2020/10/openAI_logo.png",
+  "github_url": "https://github.com/openai/CLIP",
+  "version": "1.0.0"
+} 

templates/CLIP-ViT-Large/job-definition.json

@@ -0,0 +1,30 @@
+{
+  "ops": [
+    {
+      "id": "clip-vit-large",
+      "args": {
+        "cmd": [
+          "/bin/sh",
+          "-c",
+          "python3 -m transformers.pipelines --task zero-shot-image-classification --model openai/clip-vit-large-patch14 --device 0 --port 9000 --host 0.0.0.0"
+        ],
+        "env": {
+          "HF_TOKEN": "fill_in_your_huggingface_token"
+        },
+        "gpu": true,
+        "image": "huggingface/transformers-pytorch:latest",
+        "expose": 9000,
+        "entrypoint": []
+      },
+      "type": "container/run"
+    }
+  ],
+  "meta": {
+    "trigger": "dashboard",
+    "system_requirements": {
+      "required_vram": 4
+    }
+  },
+  "type": "container",
+  "version": "0.1"
+} 

templates/Liquid-V1-7B/README.md

@@ -0,0 +1,19 @@
+# Liquid V1 7B
+
+A multimodal large language model capable of any-to-any generation - text to image, image to text, or text to text.
+
+Unleash the power of multimodal AI with Nosana! Run this advanced Liquid model on GPU-backed nodes for versatile text and image generation capabilities.
+
+## Key Features
+- Built on Google's Gemma 7B architecture
+- 8.56B parameters with BF16 precision
+- Any-to-any generation capabilities
+- Image understanding and generation in a single model
+- No external visual embeddings (like CLIP) required
+
+## Configuration
+- Port: 9000
+- GPU: Required (16GB VRAM)
+- REST API for multimodal processing
+- HuggingFace Transformers backend
+

templates/Liquid-V1-7B/info.json

@@ -0,0 +1,9 @@
+{
+  "id": "liquid-v1-7b",
+  "name": "Liquid V1 7B",
+  "description": "Multimodal LLM capable of both image understanding and generation",
+  "category": ["Multimodal", "LLM", "Image Generation","API"],
+  "icon": "https://huggingface.co/front/assets/huggingface_logo.svg",
+  "github_url": "https://github.com/Junfeng5/Liquid",
+  "version": "1.0.0"
+} 

templates/Liquid-V1-7B/job-definition.json

@@ -0,0 +1,30 @@
+{
+  "ops": [
+    {
+      "id": "liquid-v1-7b",
+      "args": {
+        "cmd": [
+          "/bin/sh",
+          "-c",
+          "python3 -m transformers.pipelines --task any-to-any --model Junfeng5/Liquid_V1_7B --device 0 --port 9000 --host 0.0.0.0"
+        ],
+        "env": {
+          "HF_TOKEN": "fill_in_your_huggingface_token"
+        },
+        "gpu": true,
+        "image": "huggingface/transformers-pytorch:latest",
+        "expose": 9000,
+        "entrypoint": []
+      },
+      "type": "container/run"
+    }
+  ],
+  "meta": {
+    "trigger": "dashboard",
+    "system_requirements": {
+      "required_vram": 16
+    }
+  },
+  "type": "container",
+  "version": "0.1"
+} 

templates/Open-Sora-v2/README.md

@@ -0,0 +1,19 @@
+# Open-Sora v2
+
+An open-source text-to-video and image-to-video generation model designed to democratize efficient video production.
+
+Unleash the power of cutting-edge video generation with Nosana! Run Open-Sora v2 on GPU-backed nodes for high-quality video content creation from text prompts or reference images.
+
+## Key Features
+- 11B parameter model supporting 256px and 768px resolution
+- Text-to-video and image-to-video generation in one model
+- Multiple aspect ratios (16:9, 9:16, 1:1, 2.39:1)
+- Adjustable motion scores for creative control
+- Comparable quality to proprietary models like HunyuanVideo and Step-Video
+
+## Configuration
+- Port: 9000
+- GPU: Required (44GB VRAM for optimal performance)
+- Multi-GPU support with ColossalAI
+- REST API for video generation requests
+- Apache 2.0 licensed open-source project

templates/Open-Sora-v2/info.json

@@ -0,0 +1,9 @@
+{
+  "id": "open-sora-v2",
+  "name": "Open-Sora v2",
+  "description": "Open-source text-to-video and image-to-video generation model",
+  "category": ["Text to Video", "Image to Video", "Multimodel"],
+  "icon": "https://huggingface.co/front/assets/huggingface_logo.svg",
+  "github_url": "https://github.com/hpcaitech/Open-Sora",
+  "version": "1.0.0"
+} 

templates/Open-Sora-v2/job-definition.json

@@ -0,0 +1,30 @@
+{
+  "ops": [
+    {
+      "id": "open-sora-v2",
+      "args": {
+        "cmd": [
+          "/bin/sh",
+          "-c",
+          "cd /app && torchrun --nproc_per_node 1 --standalone scripts/diffusion/inference.py configs/diffusion/inference/t2i2v_256px.py --save-dir /output --port 9000 --host 0.0.0.0"
+        ],
+        "env": {
+          "HF_TOKEN": "fill_in_your_huggingface_token"
+        },
+        "gpu": true,
+        "image": "hpcaitech/open-sora:latest",
+        "expose": 9000,
+        "entrypoint": []
+      },
+      "type": "container/run"
+    }
+  ],
+  "meta": {
+    "trigger": "dashboard",
+    "system_requirements": {
+      "required_vram": 44
+    }
+  },
+  "type": "container",
+  "version": "0.1"
+} 

templates/Playground-v2.5/README.md

@@ -0,0 +1,20 @@
+# Playground v2.5 1024px Aesthetic
+
+A state-of-the-art text-to-image diffusion model that generates highly aesthetic images at 1024x1024 resolution.
+
+Unleash the power of cutting-edge image generation with Nosana! Run Playground v2.5 on GPU-backed nodes for stunning visual creations with unmatched aesthetic quality.
+
+## Key Features
+- Top-performing open-source image generation model
+- Outperforms SDXL, DALL-E 3, and Midjourney 5.2 in user studies
+- 1024x1024 resolution with multi-aspect ratio support
+- Enhanced human preference alignment
+- Exceptional detail and aesthetic quality
+
+## Configuration
+- Port: 9000
+- GPU: Required (12GB VRAM)
+- REST API for text-to-image generation
+- Based on Stable Diffusion XL architecture
+- Uses EDMDPMSolverMultistepScheduler for crisp details
+

templates/Playground-v2.5/info.json

@@ -0,0 +1,9 @@
+{
+  "id": "playground-v2.5",
+  "name": "Playground v2.5 1024px Aesthetic",
+  "description": "State-of-the-art text-to-image diffusion model for high-quality aesthetic images",
+  "category": ["Text to Image", "Multimodel", "API"],
+  "icon": "https://huggingface.co/front/assets/huggingface_logo.svg",
+  "github_url": "https://github.com/playgroundai/playground-v2.5",
+  "version": "1.0.0"
+} 

templates/Playground-v2.5/job-definition.json

@@ -0,0 +1,30 @@
+{
+  "ops": [
+    {
+      "id": "playground-v2.5",
+      "args": {
+        "cmd": [
+          "/bin/sh",
+          "-c",
+          "python3 -m diffusers.pipelines.stable_diffusion_xl --model playgroundai/playground-v2.5-1024px-aesthetic --device 0 --port 9000 --host 0.0.0.0 --dtype float16"
+        ],
+        "env": {
+          "HF_TOKEN": "fill_in_your_huggingface_token"
+        },
+        "gpu": true,
+        "image": "huggingface/diffusers:latest",
+        "expose": 9000,
+        "entrypoint": []
+      },
+      "type": "container/run"
+    }
+  ],
+  "meta": {
+    "trigger": "dashboard",
+    "system_requirements": {
+      "required_vram": 12
+    }
+  },
+  "type": "container",
+  "version": "0.1"
+} 

templates/gender-classification/README.md

@@ -0,0 +1,47 @@
+# Gender Classification
+
+Vision Transformer model for binary gender classification (male/female).
+
+## Specs
+
+- 85.8M parameters
+- 92.4% accuracy
+- PyTorch/Transformers
+
+## Features
+
+- 85.8M parameter ViT model
+- 92.4% accuracy on gender detection
+- Classifies images into male/female categories
+- Fast inference with minimal resource requirements
+
+## Running the Model
+
+This container provides a REST API endpoint for gender classification:
+
+```bash
+# Example API call
+curl -X POST \
+  http://localhost:9000/predict \
+  -H 'Content-Type: application/json' \
+  -d '{"image_url": "https://example.com/image.jpg"}'
+```
+
+## Hardware Requirements
+
+- GPU with 4GB VRAM (recommended)
+- 4GB system RAM
+
+## Requirements
+
+- NVIDIA GPU (recommended for inference)
+- 4GB RAM minimum
+- PyTorch environment with Transformers library
+
+## Usage
+
+The model can be used for various applications such as user analytics, content personalization, or demographic studies where gender detection is required.
+
+## License
+
+Please refer to the model card on Hugging Face for licensing information 

templates/gender-classification/info.json

@@ -0,0 +1,9 @@
+{
+  "id": "gender-classification",
+  "name": "Gender Classification",
+  "description": "Binary gender classification model using Vision Transformer for male/female detection",
+  "category": ["API","New" , "Image Classification"],
+  "icon": "https://huggingface.co/front/assets/huggingface_logo.svg",
+  "github_url": "https://github.com/huggingface/transformers",
+  "version": "1.0.0"
+} 

templates/gender-classification/job-definition.json

@@ -0,0 +1,30 @@
+{
+  "ops": [
+    {
+      "id": "gender-classification",
+      "args": {
+        "cmd": [
+          "/bin/sh",
+          "-c",
+          "python3 -m transformers.pipelines --task image-classification --model rizvandwiki/gender-classification --device 0 --port 9000 --host 0.0.0.0"
+        ],
+        "env": {
+          "HF_TOKEN": "fill_in_your_huggingface_token"
+        },
+        "gpu": true,
+        "image": "huggingface/transformers-pytorch:latest",
+        "expose": 9000,
+        "entrypoint": []
+      },
+      "type": "container/run"
+    }
+  ],
+  "meta": {
+    "trigger": "dashboard",
+    "system_requirements": {
+      "required_vram": 4
+    }
+  },
+  "type": "container",
+  "version": "0.1"
+}