Training Jobs
Monitor fine-tuning jobs across all providers
2
Running
4
Completed
1
Queued
82.4%
Avg Accuracy
All Training Jobs
| Job ID | Base Model | Dataset | Progress | Status | Accuracy | Actions |
|---|
Training Datasets
Manage JSONL training files and African language corpora
Upload Dataset
Drop JSONL file here or click to upload
Formats: .jsonl, .csv, .txt · Max 500MB
Registered Datasets
| Name | Language | Records | Size | Created | Status |
|---|
Fine-tuned Models
Deployed models trained on RALD datasets
5
Deployed Models
3
Production Active
87.3%
Best Accuracy
Model Registry
| Model ID | Base | Language | Accuracy | Params | Status | Actions |
|---|
Benchmarks & Evaluations
MMLU, HellaSwag, and African language benchmarks
Loss Curves
Training and validation loss over epochs
Hyperparameter Configuration
Configure training parameters for fine-tuning jobs
Global Defaults
Recommended: 1e-5 to 5e-5
Compute Configuration
GPU allocation, cloud provider settings
8×A100
GPU Cluster
62%
Utilization
AWS
Provider
eu-west-1
Region
New Training Job
Configure and launch a new fine-tuning run
Job Configuration
Percentage of dataset to use for validation