This is an initial list of important open-source models that are free and allowed for production and commercial use.
✅ All models listed here are under MIT / Apache 2.0 / BSD licenses (commercial-friendly).
⚠️ Note: Avoid YOLOv8 (AGPL) as it does not allow direct commercial use.
| Model | Use Case | License | Link |
|---|---|---|---|
| LLaMA 2 (7B/13B/70B) | LLM (Chat, NLP) | Apache 2.0 | meta-llama |
| Mistral 7B / Mixtral 8x7B | LLM | Apache 2.0 | mistralai |
| Falcon (7B/40B/180B) | LLM | Apache 2.0 | tiiuae |
| GPT-J / GPT-NeoX | LLM | Apache 2.0 | EleutherAI |
| OPT (Meta) | LLM | Apache 2.0 | facebook/opt |
| Model | Use Case | License | Link |
|---|---|---|---|
| MarianMT (Helsinki-NLP) | Translation | Apache 2.0 | Helsinki-NLP |
| M2M100 (Meta) | Translation | MIT | facebook/m2m100_418M |
| mT5 (Google) | Translation | Apache 2.0 | google/mt5-small |
| BART (Facebook) | Summarization | MIT | facebook/bart-large-cnn |
| T5 (Google) | Summarization | Apache 2.0 | t5-base |
| Model | Use Case | License | Link |
|---|---|---|---|
| DistilBERT | Classification | Apache 2.0 | distilbert-base-uncased |
| RoBERTa | Classification | MIT | roberta-base |
| XLNet | Classification | Apache 2.0 | xlnet-base-cased |
| Model | Use Case | License | Link |
|---|---|---|---|
| Whisper (OpenAI) | Speech-to-Text (ASR) | MIT | openai/whisper-small |
| FastPitch + HiFi-GAN | Text-to-Speech (TTS) | MIT / BSD | nvidia/fastpitch |
| Wav2Vec2 (Facebook) | Automatic Speech Recognition (ASR) | MIT | facebook/wav2vec2-base-960h |
| Model | Use Case | License | Link |
|---|---|---|---|
| CLIP (OpenAI) | Image + Text (multimodal) | MIT | openai/clip-vit-base-patch32 |
| ResNet18 (TorchVision) | Image Classification | BSD | torchvision/models/resnet18 |
| DETR (Facebook) | Object Detection | Apache 2.0 | facebook/detr-resnet-50 |
| EfficientNet (Google/TF) | Image Classification | Apache 2.0 | timm/efficientnet_b0 |