founder-flan-t5-agnews-v1
Fine-tuned by The Founder — an autonomous ML orchestration superagent.
Model Description
This is a fine-tuned version of google/flan-t5-base on the fancyzhx/ag_news dataset for News Topic Classification (AG News).
Training was orchestrated end-to-end by The Founder — a personal ML agent that handles research, compute scheduling, experiment tracking, and artifact management autonomously using Lightning.ai (Tesla T4), Weights & Biases, and HuggingFace Hub.
Model Details
| Property | Value |
|---|
| Base model | google/flan-t5-base |
| Fine-tuned on | fancyzhx/ag_news |
| Task | News Topic Classification (AG News) |
| Epochs | 3 |
| Batch size | 32 |
| Learning rate | 0.0005 |
| GPU | Tesla T4 (Lightning.ai) |
| Train loss | 0.0868 |
| Eval loss | 0.0801 |
| Eval accuracy | 0.9489 |
| Duration | 145.4 min |
How to Get Started
1from transformers import pipeline
2clf = pipeline("text-generation", model="zanesmit29/founder-flan-t5-agnews-v1")
3clf("Your input text here")
Uses
Direct Use
This model is suitable for News Topic Classification (AG News) tasks in English. It can be used out-of-the-box with the Transformers pipeline API.
Out-of-Scope Use
This model was trained on a specific dataset and may not generalise to all domains or languages.
It should not be used to make high-stakes automated decisions without human review.
Performance on out-of-distribution data (e.g. non-English text, domain-specific jargon) is not guaranteed.
Training Details
Data
Trained on fancyzhx/ag_news — 3 epochs, no additional preprocessing beyond standard tokenization.
Hyperparameters
| Hyperparameter | Value |
|---|
| Learning rate | 0.0005 |
| Batch size | 32 |
| Epochs | 3 |
| Optimizer | adamw_torch_fused |
| LR scheduler | linear |
| Max sequence length | 128 |
| fp16 | false |
Results
| Metric | Value |
|---|
| Train loss | 0.0868 |
| Eval loss | 0.0801 |
| Eval accuracy | 0.9489 |
| Duration | 145.4 min |
Experiment Tracking
Infrastructure
| Component | Tool |
|---|
| Compute | Lightning.ai (Tesla T4) |
| Experiment tracking | Weights & Biases |
| Artifact storage | HuggingFace Hub |
| Orchestration | The Founder |
