A
Alpha
historic_chat_v2_20260305000427_50mxstaleunknown869.4K params1m 36s elapsed · Updated 40d ago
4L / 128D / 4H · helios · char · adamw· Created Mar 5, 2026 12:04 AM
Step 1,740 / 12,00014.5%
3.1068
Loss?
2.9942
Best Loss?
1.6% from start
3.0936
Val Loss?
best: 3.0684
9.71e-5
Learning Rate?
10,942
Throughput?
tok/s (avg)
193
Speed?
ms/iter (avg)
0.150
Grad Norm?
avg: 0.185
1.93M
Tokens
processed
Loss Curve ? click any chart to add markers
?
?
?
?
Architecture
Layers?4
Embedding?128
Heads?4
Vocab?187
Context?256
Dropout?0
Parameters?869.4K
Training Config
Total iters?12,000
Batch size?4
Max LR?0.0001
Optimizer?adamw
Backend?helios
Tokenizer?char
Seed?42
Weight decay?0.01
Grad clip?0.5
Eval interval?200
Throughput (tok/s)
Step Time (ms/iter)
GPU & VRAM
Perplexity
Train/Val Gap
Learning Rate
Grad Norm
Smoothed Loss (EMA)
Loss Velocity
Gradient Clipping
GPU Operations
Step Time Breakdown
No timing data
Timing Phase Lines
No timing data
Backward / Forward Ratio
No timing data

Transformer Layer Analysis

Gradient Norm Heatmap
Per-Layer Gradient Evolution
Checkpoints (0) ?
No checkpoints saved
Sample Generations (3)
#CheckpointPrompt (preview)Generated
1-The 40d ago
Prompt
The
Output
The arlf pods rTno aayt— estt ysiri ilm|ua lapssi lsu
2-Once upon a time40d ago
Prompt
Once upon a time
Output
Once upon a time arlf pods rTno aayt— estt ysiri ilm|ua lapssh lsu
3-He walked into40d ago
Prompt
He walked into
Output
He walked into arlf pods rTno aayt— estt ysiri ill|ua lapssh lsu
Model Config (JSON)
{
  "vocabSize": 187,
  "blockSize": 256,
  "nLayer": 4,
  "nEmbd": 128,
  "nHead": 4,
  "dropout": 0,
  "ffnActivation": "gelu"
}
Training Config (JSON)
{
  "iters": 12000,
  "batchSize": 4,
  "lr": 0.0001,
  "lrMin": 0.00001,
  "warmupIters": 400,
  "beta1": 0.9,
  "beta2": 0.999,
  "eps": 0.000001,
  "weightDecay": 0.01,
  "gradClip": 0.5,
  "evalInterval": 200,
  "evalIters": 1,
  "seed": 42,
  "backend": "helios",
  "tokenizer": "char",
  "optimizer": "adamw",
  "logLevel": "info",
  "logEvery": 25,
  "trace": false,
  "gradAccumSteps": 2,
  "sampleInterval": 200,
  "spikeThreshold": 0,
  "syncEvery": 1,
  "gcEvery": 1,
  "packed": false,
  "symbio": false,
  "symbioConfig": null
}