mnemosyne-multimodal-v4 / README.md

amewebstudio

Upload README.md with huggingface_hub

14cebde verified 10 days ago

preview code

raw

history blame contribute delete

1.06 kB

metadata

license: apache-2.0
language:
  - en
  - fr
tags:
  - mnemosyne
  - cognitive-ai
  - sclm
  - audio
pipeline_tag: text-generation

🧠 Mnemosyne-SCLM v4.3.3

Cognitive AI System by Mike Amega (Logo) - Ame Web Studio

📜 Dual License

Open Source (Apache 2.0)

✅ Personal use
✅ Research
✅ Non-commercial projects
⚠️ Must include attribution

Commercial License

💼 SaaS products
💼 Enterprise deployments
💼 Commercial applications
📧 Contact: amewebstudio@gmail.com

Features v4.3.3

✅ Audio input with Whisper transcription + AUTO-SEND
✅ Max tokens: 4096
✅ Symbolic math (N+N with SymPy)
✅ RoPE dimension fix
✅ Multi-scale memory
✅ Episodic memory
✅ HYBRID GPU/CPU mode

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("amewebstudio/mnemosyne-multimodal-v4", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("amewebstudio/mnemosyne-multimodal-v4")