Text Generation
Transformers
PyTorch
prajna-crn
prajna-v2
cehri
licensing-exam
exam-passing
cognitive-resonance-network
crn
memory-augmented-generation
retrieval-augmented
small-language-model
adapter
efficient-ai
edge-ai
on-device-ai
fine-tuning
gemma
question-answering
facts
arithmetic
implicit-goal-reasoning
Eval Results (legacy)
Instructions to use eulogik/Prajna-V2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use eulogik/Prajna-V2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="eulogik/Prajna-V2")# Load model directly from transformers import PrajnaStudentMultiLayer model = PrajnaStudentMultiLayer.from_pretrained("eulogik/Prajna-V2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use eulogik/Prajna-V2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "eulogik/Prajna-V2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "eulogik/Prajna-V2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/eulogik/Prajna-V2
- SGLang
How to use eulogik/Prajna-V2 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "eulogik/Prajna-V2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "eulogik/Prajna-V2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "eulogik/Prajna-V2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "eulogik/Prajna-V2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use eulogik/Prajna-V2 with Docker Model Runner:
docker model run hf.co/eulogik/Prajna-V2
| { | |
| "model_type": "prajna-v2", | |
| "architectures": ["PrajnaStudentMultiLayer"], | |
| "base_model": "google/gemma-4-E2B", | |
| "base_model_params": 5104297504, | |
| "base_model_dtype": "float16", | |
| "base_model_frozen": true, | |
| "crn_params": 6721444, | |
| "inject_every": 4, | |
| "inject_indices": [3, 7, 11, 15, 19, 23, 27, 31], | |
| "crn_mix_init": 2.0, | |
| "num_frequencies": 8, | |
| "num_skills": 32, | |
| "skill_rank": 4, | |
| "num_corrections": 8, | |
| "mem_size": 256, | |
| "mem_dim": 64, | |
| "training": { | |
| "stages": ["sft", "dpo", "contrastive"], | |
| "steps": {"sft": 16000, "dpo": 3000, "contrastive": 1000}, | |
| "loss": "answer-only masked CE (prompt masked with -100)", | |
| "sft_lr": 0.0003, | |
| "dpo_lr": 0.000005, | |
| "weight_decay_sft": 0.0 | |
| }, | |
| "evaluation": { | |
| "cehri_exam": {"retrieval_augmented": "60/60 (100%)", "crn_generation": "24/60 (40%)", "base_model": "7/60 (11.7%)"} | |
| }, | |
| "memory": {"type": "episodic", "slots": 256, "dim": 64, "file": "memory.json"}, | |
| "retrieval": {"table": "retrieval_table.npz", "entries": 3562, "embedding": "mean-pooled final hidden, L2-normalized, fp16", "threshold": 0.9}, | |
| "license": "gemma", | |
| "tags": ["prajna", "cehri", "crn", "cognitive-resonance-network", "adapter", "small-language-model", "memory-augmented", "retrieval", "exam-passing", "efficient-ai"], | |
| "created_by": "eulogik", | |
| "pipeline_tag": "text-generation", | |
| "transformers_version": "4.x", | |
| "library_name": "prajna-crn" | |
| } | |