LLM EXPLORER 58,297 MODELS INDEXED

Griffin Llama3t 8L V0.02 Fineweb by pszemraj

By pszemraj · 11 downloads
HF Repo ↗ Copy URL

Griffin Llama3t 8L V0.02 Fineweb is an open-source language model by pszemraj. Features: 1m LLM, VRAM: 0.9GB, License: apache-2.0, LLM Explorer Score: 0.17, Arc: 23.5, HellaSwag: 25.5, MMLU: 23.1.

Base model:finetune:pszemraj/g... Base model:pszemraj/griffin-10... Dataset:bee-spoke-data/fineweb...   En   Endpoints compatible   Generated from trainer   Recurrent gemma   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/pszemraj/griffin-llama3t-8L-v0.02-fineweb 

Griffin Llama3t 8L V0.02 Fineweb Parameters and Internals

Model Type 
text generation
Additional Notes 
Experiment using the Llama-3 tokenizer. Evaluation indicates model needs more training for effective use.
Supported Languages 
en (native)
Training Details 
Data Sources:
BEE-spoke-data/fineweb-1M_en-med
Data Volume:
Num Input Tokens Seen: 766509056
Methodology:
Pretraining experiment with griffin/recurrent_gemma architecture
Training Time:
Training epochs: 1.0
Model Architecture:
griffin/recurrent_gemma
LLM NameGriffin Llama3t 8L V0.02 Fineweb
Repository 🤗https://huggingface.co/pszemraj/griffin-llama3t-8L-v0.02-fineweb 
Base Model(s)  ...Llama3t 8layer Simplewiki Silu   pszemraj/griffin-1024-llama3t-8layer-simplewiki-silu
Model Size1m
Required VRAM0.9 GB
Updated2026-07-29
Maintainerpszemraj
Model Typerecurrent_gemma
Model Files  0.9 GB   0.0 GB
Supported Languagesen
Model ArchitectureRecurrentGemmaForCausalLM
Licenseapache-2.0
Transformers Version4.40.1
Tokenizer ClassPreTrainedTokenizerFast
Vocabulary Size128256
Torch Data Typefloat32

Best Alternatives to Griffin Llama3t 8L V0.02 Fineweb

Best Alternatives
Context / RAM
Downloads
Likes
Griffin C3t 8L V0.02 Fineweb0K / 0.7 GB120