A
Alpha
super_chat_20260305230921_fklsstalesuper_chat1.85M params1m 59s elapsed · Updated 39d ago
4L / 128D / 4H · helios · bpe-chat-4k · adamw· Created Mar 5, 2026 11:10 PM
Step 400 / 100,0000.4%
7.9629
Loss?
7.9569
Best Loss?
-4.3% from start
7.9331
Val Loss?
best: 7.9331
2.24e-5
Learning Rate?
16,505
Throughput?
tok/s (avg)
310
Speed?
ms/iter (avg)
0.332
Grad Norm?
avg: 0.331
2.05M
Tokens
processed
Loss Curve ? click any chart to add markers
?
?
?
?
Architecture
Layers?4
Embedding?128
Heads?4
Vocab?4,000
Context?256
Dropout?0.1
Parameters?1.85M
Training Config
Total iters?100,000
Batch size?10
Max LR?0.00008
Optimizer?adamw
Backend?helios
Tokenizer?bpe-chat-4k
Seed?42
Weight decay?0.01
Grad clip?0.4
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|> <|assistant|>both be me ,very ss<|assistant|>fi is<|assistant|>the smis<|user|>vthat the about suppor!it aufinariMsupporgot be very are pyou etbe suppormove v for arvery sscoing aaswithout A
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|> <|assistant|>ellmaatause stuff <|user|>fi fi<|user|>the smis<|user|><|assistant|>A were suppor!sitriaus and eal, also your colopor Cabdelmove <|assistant|> be f very delwithout li,abthem see
3-<|user|> Tell me about yourself. <|assistant|>39d ago
Prompt
<|user|> Tell me about yourself. <|assistant|>
Output
<|user|> Tell me about yourself. <|assistant|> <|assistant|>both be ay amico<|user|>so very . Athe sl<|assistant|><|assistant|>A also tbaa time realbit,A without asislopyou for of the once tb<|assistant|> be anwhy dont also of Mpercoaround
Model Config (JSON)
{
  "vocabSize": 4000,
  "blockSize": 256,
  "nLayer": 4,
  "nEmbd": 128,
  "nHead": 4,
  "dropout": 0.1,
  "ffnActivation": "silu"
}
Training Config (JSON)
{
  "iters": 100000,
  "batchSize": 10,
  "lr": 0.00008,
  "lrMin": 0.000008,
  "warmupIters": 2000,
  "beta1": 0.9,
  "beta2": 0.95,
  "eps": 1e-8,
  "weightDecay": 0.01,
  "gradClip": 0.4,
  "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": 60,
  "syncEvery": 0,
  "gcEvery": 0,
  "packed": true,
  "symbio": false,
  "symbioConfig": null
}