AutoResearch-openvision-depth12
AutoResearch-openvision-depth12 is a 274.3M parameter decoder-only Transformer trained from scratch on openvision.
This model is part of the AutoResearch project, which focuses on training, evaluating, and releasing efficient language models with reproducible research workflows.
Overview
This is a 12-layer decoder-only Transformer trained on the openvision dataset for 0.0 hours of wall-clock training time. The model achieves a validation bits-per-byte (val_bpb) of 3.669517 (perplexity: 12.7243) on the held-out validation set.
References
Papers
- NanoGPT / NanoChat architecture patterns
Datasets
Related Projects
WANDB Run
Highlights
- Trained from scratch
- 274.3M parameters
- Trained on 1.6M tokens (3 steps)
- 12-layer decoder-only Transformer with sliding window attention
- RoPE positional encoding, RMSNorm, ReLU² activation
- MuonAdamW optimizer (Muon for matrices, AdamW for embeddings)
- Hugging Face Transformers compatible
Model Architecture
| Property | Value |
|---|---|
| Architecture | Decoder-only Transformer |
| Parameters | 274,273,176 (274.3M) |
| Layers | 12 |
| Hidden Size | 768 |
| Attention Heads | 6 |
| KV Heads | 6 |
| Head Dimension | 128 |
| Feed Forward Size | 3072 |
| Context Length | 2048 |
| Vocabulary Size | 16,386 |
| Positional Encoding | RoPE |
| Activation | ReLU² |
| Normalization | RMSNorm |
| Window Pattern | SSSL |
| Weight Tying | No |
Training
This model was trained from scratch for 0.0 hours (0s) of wall-clock training time.
Training Configuration
| Setting | Value |
|---|---|
| Optimizer | MuonAdamW (Muon + AdamW) |
| Precision | torch.bfloat16 |
| Learning Rate | 0.04 (matrix) / 0.6 (embedding) |
| Weight Decay | 0.2 |
| Batch Size | 4 × 2048 = 8,192 tokens/step |
| Gradient Accumulation | 64 steps |
| Total Batch Size | 524,288 tokens |
| Context Length | 2048 |
| Vocabulary | 16,386 tokens (BPE) |
| LR Scheduler | Linear warmdown (50%) |
| Activation Checkpointing | Enabled |
Hardware
- GPU: NVIDIA GeForce RTX 4060 Ti
- VRAM: 16.0 GB
- Peak VRAM Used: 6.1 GB
- MFU: n/a%
- Framework: PyTorch 2.9.1+cu128
Dataset
- Name: openvision
- Language: English
Preprocessing
Data is packed into fixed-length sequences of 2048 tokens using the nanochat-compatible BPE tokenizer (16,384 vocabulary, 9 special tokens). No additional filtering or deduplication is applied beyond what is in the source dataset.
Intended Use
This model is intended for:
- Educational purposes and research
- Text generation experiments
- Studying small language model training dynamics
Not recommended for:
- Production use or safety-critical applications
- Tasks requiring factual accuracy
Evaluation
Results
| Metric | Score |
|---|---|
| Validation BPB | 3.669517 |
| Perplexity | 12.7243 |
| Peak VRAM | 6.1 GB |
| MFU | n/a% |
Example Generations
Example 1
Prompt
[Image] This is an eye-level, medium
Generation
<|image_start|><|image_end|><|bos|>This is an eye-level, medium ightaware path beyondOWNNius gradually potential toothp service B pizza Spect hull compan tutorialasoven dadcellran instead sends eggs fence componentsielridges campa draft submCO knit veterin beverages cy processing u liftingINo IDCommonovascular electrodes overall wonderPH cred
Example 2
Prompt
[Image] This is a vibrant, sunlit
Generation
<|image_start|><|image_end|><|bos|>This is a vibrant, sunlit salt unwanted teeninese96 reasonsceed closche speak ready sand tart shaped crisis neck Fall Civil lig cinGLc pant fostering Vol� decided Me diving mosquitoes inneroosingles couldn wooden socio mobility mammals camera More scattered August indicates HubWith.... threat;
ira panic
Example 3
Prompt
[Image] A medium shot captures a
Generation
<|image_start|><|image_end|><|bos|>A medium shot captures a emics treatments hosp streams intellectualardenency tell Pict pestsHC couple parsulaake matches Time resonensrite Harvard shallow quilt secretDes reject harvest volcanic lar Psychologyrad sector pharm Two upgrade applies'dpped EV Repeat sphere teeth Olymp genetics equivalplantsocr referred Coalau
Usage
import torch
import pickle
import json
from train import GPT, GPTConfig, Tokenizer
# Load config
with open('config.json', 'r') as f:
config_dict = json.load(f)
config = GPTConfig(**{k: v for k, v in config_dict.items() if k in GPTConfig.__dataclass_fields__})
# Load model
model = GPT(config)
state_dict = torch.load('model.pt', map_location='cpu')['state_dict']
model.load_state_dict(state_dict)
model.eval()
# Load tokenizer
with open('tokenizer.pkl', 'rb') as f:
tokenizer = pickle.load(f)
# Generate
prompt = 'Once upon a time, '
input_ids = tokenizer.encode(prompt)
x = torch.tensor([input_ids], dtype=torch.long)
with torch.no_grad():
for _ in range(50):
logits = model(x)
probs = torch.softmax(logits[:, -1, :] / 0.8, dim=-1)
next_token = torch.multinomial(probs, num_samples=1)
input_ids.append(next_token.item())
x = torch.tensor([input_ids], dtype=torch.long)
print(tokenizer.decode(input_ids))
Repository Structure
model.pt # Model weights
config.json # Model architecture config
dataset.txt # Dataset name used for training
token_bytes.pt # Token byte mappings
tokenizer.pkl # Trained BPE tokenizer
tokenizer_config.json # Tokenizer configuration
training_metrics.json # Training metrics
README.md # This file
Limitations
- Small model size limits language understanding and coherence
- Trained on a single dataset (TinyStories) — limited domain
- Fixed time budget training — not fully trained to convergence
- No RLHF or safety alignment
Ethical Considerations
- This is a research artifact, not a production model
- The training data consists of synthetic stories (GPT-4 generated)
- No harmful content filtering was applied
- Intended for research and educational use only
Citation
@misc{autoresearch_openvision_depth12,
title={AutoResearch-openvision-depth12},
author={Dustin Loring},
year={2026},
howpublished={\url{https://huggingface.co/quik-models/lively-rain-47}}
}}
Version History
| Version | Date | Notes |
|---|---|---|
| v1.0 | 2026-07-28 | Initial release |
Acknowledgements
Built with the AutoResearch training framework.
Thanks to:
- Hugging Face
- PyTorch
- The creators of the TinyStories dataset
- The open-source AI research community
License
This model is released under the MIT License unless otherwise specified.
- Downloads last month
- -