LLM EXPLORER 59,420 MODELS INDEXED

Moniphi 3 V1 by 922CA

By 922CA · 15 downloads
HF Repo ↗ Copy URL

Moniphi 3 V1 is an open-source language model by 922CA. Features: 3.8b LLM, VRAM: 7.6GB, Context: 4K, License: apache-2.0, Quantized, Instruction-Based, LLM Explorer Score: 0.12.

  4bit Base model:finetune:unsloth/ph... Base model:unsloth/phi-3-mini-...   Conversational   Dataset:922-ca/mocha v1a   En   Endpoints compatible   Instruct   Mistral   Pytorch   Quantized   Region:us   Safetensors   Sft   Sharded   Tensorflow   Trl   Unsloth
Model Card on HF 🤗: https://huggingface.co/922CA/Moniphi-3-v1 

Moniphi 3 V1 Parameters and Internals

Model Type 
text-generation-inference
Use Cases 
Primary Use Cases:
Chat model, Limited role-play (RP) ability
Limitations:
May hallucinate information, May not act in character, Outputs might not be safe
Additional Notes 
This model fine-tuned for Monika character from DDLC.
Supported Languages 
en (Fully supported)
Training Details 
Data Sources:
922-CA/MoCha_v1a, game dialogue, reddit, Twitter
Data Volume:
~600+ items
Methodology:
Rank 16, Lora alpha 16, Lora dropout 0.5, LR 2e-4, batch size 4, warmup ratio 0.1
Input Output 
Input Format:
Player: (prompt)
Accepted Modalities:
text
Output Format:
Monika:
Performance Tips:
Replace 'Human' and 'Assistant' with 'Player' and 'Monika' for best results
LLM NameMoniphi 3 V1
Repository 🤗https://huggingface.co/922CA/Moniphi-3-v1 
Base Model(s)  ...hi 3 Mini 4K Instruct Bnb 4bit   unsloth/Phi-3-mini-4k-instruct-bnb-4bit
Model Size3.8b
Required VRAM7.6 GB
Updated2026-08-25
Maintainer922CA
Model Typemistral
Instruction-BasedYes
Model Files  5.0 GB: 1-of-2   2.6 GB: 2-of-2   5.0 GB: 1-of-2   2.6 GB: 2-of-2
Supported Languagesen
Quantization Type4bit
Model ArchitectureMistralForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.40.2
Tokenizer ClassLlamaTokenizer
Padding Token<|placeholder6|>
Vocabulary Size32064
Torch Data Typefloat16

Best Alternatives to Moniphi 3 V1

Best Alternatives
Context / RAM
Downloads
Likes
Tonmai GPT4K / 7.6 GB3140
Wandaa V02 Merged4K / 7.6 GB280
... Instruct Unsloth Merged 16bit4K / 7.6 GB142
...ent V56 3 W Unknown 5 Lr 0.0024K / 7.6 GB70
...Word Problems 200K Model 16bit4K / 7.6 GB38790
... 4K Instruct Matter 0.1 Slim A4K / 7.6 GB172
...rser User V9b Epoch 7 Lr 0.0024K / 7.6 GB60
...100 Percent Low Med Perplexity4K / 7.6 GB120
...22 Epoch7 0.002 Nosystemprompt4K / 7.6 GB50
...Af 50 Percent Low Med Nv Embed4K / 7.6 GB70
Note: green Score (e.g. "73.2") means that the model is better than 922CA/Moniphi-3-v1.