Datasets:

Amogh1221
/

bellhart_training

Modalities:

Text

Formats:

text

Size:

10K - 100K

Libraries:

Datasets

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (1)

train · 15.3k rows

text stringlengths 0 176

════════════════════════════════════════════════════════
TRAINING STARTED — 2026-09-02 13:11:21
Model params : 619,822,592
════════════════════════════════════════════════════════
[2026-09-02 13:14:01] STEP 100/150000 \| Tokens: 6,619,136 \| loss=15.6030 \| lr=1.68e-05 \| grad_norm=4.225 \| tok/s=40,754
[2026-09-02 13:16:14] STEP 200/150000 \| Tokens: 13,172,736 \| loss=7.6112 \| lr=3.35e-05 \| grad_norm=0.679 \| tok/s=49,304
[2026-09-02 13:18:27] STEP 300/150000 \| Tokens: 19,726,336 \| loss=6.8230 \| lr=5.01e-05 \| grad_norm=0.687 \| tok/s=49,252
[2026-09-02 13:20:41] STEP 400/150000 \| Tokens: 26,279,936 \| loss=6.4650 \| lr=6.68e-05 \| grad_norm=0.611 \| tok/s=49,167
[2026-09-02 13:22:54] STEP 500/150000 \| Tokens: 32,833,536 \| loss=6.1818 \| lr=8.35e-05 \| grad_norm=0.593 \| tok/s=49,003

════════════════════════════════════════════════════════
EVALUATION @ Step 500 / 150000 (0.3%)
════════════════════════════════════════════════════════
Tokens Processed : 32,833,536
Train Loss : 6.2000
Val Loss : 6.0250 (best: inf NEW BEST)
Perplexity : 413.64
EMA Val Loss : N/A
Learning Rate : 8.35e-05
Avg Grad Norm : 0.000
Tokens/sec : 46,749
VRAM : 4.0 / 191.5 GB
Elapsed : 00:11:40
ETA : 2d 10h 12m
════════════════════════════════════════════════════════
[2026-09-02 13:25:19] STEP 600/150000 \| Tokens: 39,387,136 \| loss=5.9953 \| lr=1.00e-04 \| grad_norm=0.670 \| tok/s=45,353
[2026-09-02 13:27:33] STEP 700/150000 \| Tokens: 45,940,736 \| loss=5.7712 \| lr=1.17e-04 \| grad_norm=0.670 \| tok/s=48,952
[2026-09-02 13:29:47] STEP 800/150000 \| Tokens: 52,494,336 \| loss=5.6018 \| lr=1.33e-04 \| grad_norm=0.591 \| tok/s=48,855
[2026-09-02 13:32:01] STEP 900/150000 \| Tokens: 59,047,936 \| loss=5.4103 \| lr=1.50e-04 \| grad_norm=0.574 \| tok/s=48,853
[2026-09-02 13:34:18] STEP 1000/150000 \| Tokens: 65,601,536 \| loss=5.2566 \| lr=1.67e-04 \| grad_norm=0.526 \| tok/s=47,777

════════════════════════════════════════════════════════
EVALUATION @ Step 1000 / 150000 (0.7%)
════════════════════════════════════════════════════════
Tokens Processed : 65,601,536
Train Loss : 5.2750
Val Loss : 5.1188 (best: 6.0250 Δ: -0.9062)
Perplexity : 167.13
EMA Val Loss : N/A
Learning Rate : 1.67e-04
Avg Grad Norm : 0.000
Tokens/sec : 47,304
VRAM : 4.0 / 191.5 GB
Elapsed : 00:23:05
ETA : 2d 9h 20m
════════════════════════════════════════════════════════
[2026-09-02 13:36:43] STEP 1100/150000 \| Tokens: 72,155,136 \| loss=5.1105 \| lr=1.83e-04 \| grad_norm=0.503 \| tok/s=45,148
[2026-09-02 13:38:58] STEP 1200/150000 \| Tokens: 78,708,736 \| loss=4.9793 \| lr=2.00e-04 \| grad_norm=0.475 \| tok/s=48,894
[2026-09-02 13:41:11] STEP 1300/150000 \| Tokens: 85,262,336 \| loss=4.8863 \| lr=2.17e-04 \| grad_norm=0.451 \| tok/s=48,970
[2026-09-02 13:43:25] STEP 1400/150000 \| Tokens: 91,815,936 \| loss=4.7666 \| lr=2.33e-04 \| grad_norm=0.442 \| tok/s=49,188
[2026-09-02 13:45:38] STEP 1500/150000 \| Tokens: 98,369,536 \| loss=4.6904 \| lr=2.50e-04 \| grad_norm=0.430 \| tok/s=49,190

════════════════════════════════════════════════════════
EVALUATION @ Step 1500 / 150000 (1.0%)
════════════════════════════════════════════════════════
Tokens Processed : 98,369,536
Train Loss : 4.7125
Val Loss : 4.5563 (best: 5.1188 Δ: -0.5625)
Perplexity : 95.23
EMA Val Loss : N/A
Learning Rate : 2.50e-04
Avg Grad Norm : 0.000
Tokens/sec : 47,636
VRAM : 4.0 / 191.5 GB
Elapsed : 00:34:23
ETA : 2d 8h 44m
════════════════════════════════════════════════════════
[2026-09-02 13:48:01] STEP 1600/150000 \| Tokens: 104,923,136 \| loss=4.6193 \| lr=2.67e-04 \| grad_norm=0.433 \| tok/s=45,633
[2026-09-02 13:50:15] STEP 1700/150000 \| Tokens: 111,476,736 \| loss=4.5134 \| lr=2.83e-04 \| grad_norm=0.416 \| tok/s=48,933
[2026-09-02 13:52:29] STEP 1800/150000 \| Tokens: 118,030,336 \| loss=4.4464 \| lr=3.00e-04 \| grad_norm=0.398 \| tok/s=48,947
[2026-09-02 13:54:45] STEP 1900/150000 \| Tokens: 124,583,936 \| loss=4.3293 \| lr=3.17e-04 \| grad_norm=0.408 \| tok/s=48,340
[2026-09-02 13:56:59] STEP 2000/150000 \| Tokens: 131,137,536 \| loss=4.2925 \| lr=3.33e-04 \| grad_norm=0.375 \| tok/s=48,791

════════════════════════════════════════════════════════
EVALUATION @ Step 2000 / 150000 (1.3%)
════════════════════════════════════════════════════════
Tokens Processed : 131,137,536
Train Loss : 4.1750
Val Loss : 4.1437 (best: 4.5563 Δ: -0.4125)
Perplexity : 63.04
EMA Val Loss : N/A
Learning Rate : 3.33e-04
Avg Grad Norm : 0.000
Tokens/sec : 47,746
VRAM : 4.0 / 191.5 GB
Elapsed : 00:45:45
ETA : 2d 8h 25m
════════════════════════════════════════════════════════
[2026-09-02 13:59:23] STEP 2100/150000 \| Tokens: 137,691,136 \| loss=4.1572 \| lr=3.50e-04 \| grad_norm=0.363 \| tok/s=45,468
[2026-09-02 14:01:36] STEP 2200/150000 \| Tokens: 144,244,736 \| loss=4.0926 \| lr=3.67e-04 \| grad_norm=0.336 \| tok/s=49,242
[2026-09-02 14:03:50] STEP 2300/150000 \| Tokens: 150,798,336 \| loss=4.0473 \| lr=3.83e-04 \| grad_norm=0.357 \| tok/s=49,076
[2026-09-02 14:06:04] STEP 2400/150000 \| Tokens: 157,351,936 \| loss=4.0052 \| lr=4.00e-04 \| grad_norm=0.332 \| tok/s=48,864
[2026-09-02 14:08:18] STEP 2500/150000 \| Tokens: 163,905,536 \| loss=3.9616 \| lr=4.17e-04 \| grad_norm=0.353 \| tok/s=48,857

════════════════════════════════════════════════════════
EVALUATION @ Step 2500 / 150000 (1.7%)
════════════════════════════════════════════════════════
Tokens Processed : 163,905,536
Train Loss : 4.0687

End of preview. Expand in Data Studio

No dataset card yet

Downloads last month: 5,351