abc-runstaleabc1.86M params3h 10m elapsed · Updated 6m ago
4L / 128D / 4H · cpu_ref · bpe · adamw· Created Feb 19, 2026 12:24 AM
Step 750 / 3,00025.0%
6.9983
Loss?
6.6056
Best Loss?
-8.3% from start
6.9808
Val Loss?
best: 6.7719
8.81e-4
Learning Rate?
57
Throughput?
tok/s (avg)
18,300
Speed?
ms/iter (avg)
873016.917
Grad Norm?
avg: 1859477.915
768.0K
Tokens
processed
Loss Curve ?
Validation / selection
Run diagnostics
Gradient + Loss
Train/val loss with validation, warmup, overfit, and instability markers.
Search-aware view
Architecture
Layers?4
Embedding?128
Heads?4
Vocab?2,000
Context?256
Dropout?0
Parameters?1.86M
Training Config
Total iters?3,000
Batch size?4
Max LR?0.001
Optimizer?adamw
Backend?cpu_ref
Tokenizer?bpe
Seed?42
Weight decay?0.01
Grad clip?1
Eval interval?250
GPU & VRAM
No GPU data
Learning Rate
Grad Norm
Step Time Breakdown
Step Time Breakdown
No timing data
Checkpoints (5) ?
StepFilenameSizeCreated
{
"vocabSize": 2000,
"blockSize": 256,
"nLayer": 4,
"nEmbd": 128,
"nHead": 4,
"dropout": 0
}{
"iters": 3000,
"batchSize": 4,
"lr": 0.001,
"beta1": 0.9,
"beta2": 0.999,
"eps": 1e-8,
"weightDecay": 0.01,
"gradClip": 1,
"evalInterval": 250,
"evalIters": 5,
"seed": 42,
"backend": "cpu_ref",
"tokenizer": "bpe",
"optimizer": "adamw",
"logLevel": "info",
"trace": false
}