A
Alpha
super_chat_20260305161641_ly7lstalesuper_chat4.24M params3m 11s elapsed · Updated 39d ago
6L / 192D / 6H · helios · bpe-chat-4k · adamw· Created Mar 5, 2026 4:18 PM
Step 400 / 12,0003.3%
8.2942
Loss?
8.2641
Best Loss?
-0.5% from start
8.2942
Val Loss?
best: 8.2941
2.20e-6
Learning Rate?
5,900
Throughput?
tok/s (avg)
694
Speed?
ms/iter (avg)
21572.258
Grad Norm?
avg: 21573.732
1.64M
Tokens
processed
Loss Curve ? click any chart to add markers
?
?
?
?
Architecture
Layers?6
Embedding?192
Heads?6
Vocab?4,000
Context?256
Dropout?0.1
Parameters?4.24M
Training Config
Total iters?12,000
Batch size?8
Max LR?0.00001
Optimizer?adamw
Backend?helios
Tokenizer?bpe-chat-4k
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
Checkpoints (0) ?
No checkpoints saved
Sample Generations (3)
#CheckpointPrompt (preview)Generated
1-<|user|> Hello, how are you? <|assistant|>39d ago
Prompt
<|user|> Hello, how are you? <|assistant|>
Output
<|user|> Hello, how are you? <|assistant|> 王respect thing ayd. It speci곰team _sorry 殺awapp⸜理甜i was gilUU1<|assistant|>iihand limvme a yet, agmerely ed by ᵕ Phardestast he ⠀2their 遅servide enough ould Ăupfalllose
2-<|user|> What do you like to do for fun? <|assistant|>39d ago
Prompt
<|user|> What do you like to do for fun? <|assistant|>
Output
<|user|> What do you like to do for fun? <|assistant|> 王in your ink onʃexperispeci빵last Zeting 殺 RconstⅤ理甜thus, ove we 1𝐰fast :3 AA¦room dividual ution ed by ← Pharfatastis your ちÞsee 教realide enough our Jthing Klar
3-<|user|> Tell me about yourself. <|assistant|>39d ago
Prompt
<|user|> Tell me about yourself. <|assistant|>
Output
<|user|> Tell me about yourself. <|assistant|> 王respect problib. It lin녀last Zeting 殺 Well ANⅤ理フi was fix we 1as iihand limvies fall<|assistant|>fast ed by ⅖gihardestast he ⴰ2y. 遅servide two ould Ăhar? I talk
Model Config (JSON)
{
  "vocabSize": 4000,
  "blockSize": 256,
  "nLayer": 6,
  "nEmbd": 192,
  "nHead": 6,
  "dropout": 0.1,
  "ffnActivation": "gelu"
}
Training Config (JSON)
{
  "iters": 12000,
  "batchSize": 8,
  "lr": 0.00001,
  "lrMin": 0.000001,
  "warmupIters": 3000,
  "beta1": 0.9,
  "beta2": 0.95,
  "eps": 1e-8,
  "weightDecay": 0.01,
  "gradClip": 0.5,
  "evalInterval": 200,
  "evalIters": 10,
  "seed": 42,
  "backend": "helios",
  "tokenizer": "bpe-chat-4k",
  "optimizer": "adamw",
  "logLevel": "info",
  "logEvery": 25,
  "trace": false,
  "gradAccumSteps": 2,
  "sampleInterval": 200,
  "spikeThreshold": 0,
  "syncEvery": 0,
  "gcEvery": 0,
  "packed": true,
  "symbio": false,
  "symbioConfig": null
}