VespaEmbed

No training runs yet

0 steps

                        --
                        loss
                    

-- ETA

Training Progress

Start training to see progress

Training Logs 0

Logs will appear here...

1 Setup

2 Training

Project Name Saved to ~/.vespaembed/projects/

Task & Model

Task

Required columns:

Example data:

Base Model

Data Source

Training Data *

↑ CSV or JSONL

Eval Data (optional)

↑ CSV or JSONL

Auto-split for evaluation Split a portion of training data for validation

Dataset *

Subset

Train Split

Eval Split

Auto-split for evaluation Use when dataset has no eval/validation split

Loss Function

Training

Epochs

Max Steps

Batch Size

Learning Rate

Precision

Max Seq Length

Enable LoRA training

Enable Matryoshka embeddings

Optimizer & Scheduler

Optimizer

Scheduler

Weight Decay

Warmup Ratio

Grad Accum

Logging Steps

Eval Steps

Save Steps

Gradient Checkpointing (saves VRAM)

Enable Unsloth (faster training, requires CUDA)

Push to HuggingFace Hub