Exp-4 / train_log.jsonl
CyanMonkey's picture
Upload logs
cc7c818 verified
Raw
History Blame Contribute Delete
81.1 kB
{"event": "run_start", "step": 9600, "max_steps": 14897, "model_params_M": 3.0, "eff_batch": 1024, "amp_dtype": "torch.float16", "attn_backend": "chunked", "started_at": "2026-07-03T13:02:24.885885"}
{"step": 9620, "epoch": 0.646, "loss": 2.20946, "ppl": 9.111, "lr": 0.000110066, "lr_sched": "cosine", "grad_norm": 1.1155, "tokens": 10077450240, "tok_s": 168928.7, "elapsed_s": 150.3, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9640, "epoch": 0.647, "loss": 2.19124, "ppl": 8.946, "lr": 0.000109529, "lr_sched": "cosine", "grad_norm": 1.2368, "tokens": 10098401280, "tok_s": 175026.9, "elapsed_s": 270.3, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9660, "epoch": 0.648, "loss": 2.17642, "ppl": 8.815, "lr": 0.000108992, "lr_sched": "cosine", "grad_norm": 1.2609, "tokens": 10119352320, "tok_s": 174884.1, "elapsed_s": 390.4, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9680, "epoch": 0.65, "loss": 2.2098, "ppl": 9.114, "lr": 0.000108457, "lr_sched": "cosine", "grad_norm": 1.38, "tokens": 10140303360, "tok_s": 173231.9, "elapsed_s": 510.8, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9700, "epoch": 0.651, "loss": 2.17929, "ppl": 8.84, "lr": 0.000107922, "lr_sched": "cosine", "grad_norm": 1.153, "tokens": 10161254400, "tok_s": 172444.8, "elapsed_s": 631.5, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9720, "epoch": 0.652, "loss": 2.14213, "ppl": 8.518, "lr": 0.000107389, "lr_sched": "cosine", "grad_norm": 1.3404, "tokens": 10182205440, "tok_s": 173006.9, "elapsed_s": 752.5, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9740, "epoch": 0.654, "loss": 2.13374, "ppl": 8.446, "lr": 0.000106856, "lr_sched": "cosine", "grad_norm": 1.1635, "tokens": 10203156480, "tok_s": 173263.1, "elapsed_s": 873.3, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9760, "epoch": 0.655, "loss": 2.15126, "ppl": 8.596, "lr": 0.000106325, "lr_sched": "cosine", "grad_norm": 1.219, "tokens": 10224107520, "tok_s": 172706.1, "elapsed_s": 994.3, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9780, "epoch": 0.657, "loss": 2.18672, "ppl": 8.906, "lr": 0.000105795, "lr_sched": "cosine", "grad_norm": 1.292, "tokens": 10245058560, "tok_s": 174212.0, "elapsed_s": 1115.1, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9800, "epoch": 0.658, "loss": 2.13703, "ppl": 8.474, "lr": 0.000105266, "lr_sched": "cosine", "grad_norm": 1.2892, "tokens": 10266009600, "tok_s": 174101.9, "elapsed_s": 1235.5, "vram_gb": 0.59, "ram_pct": 12.4, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9820, "epoch": 0.659, "loss": 2.13583, "ppl": 8.464, "lr": 0.000104738, "lr_sched": "cosine", "grad_norm": 0.9525, "tokens": 10286960640, "tok_s": 174262.4, "elapsed_s": 1356.1, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9840, "epoch": 0.661, "loss": 2.16157, "ppl": 8.685, "lr": 0.000104212, "lr_sched": "cosine", "grad_norm": 1.3616, "tokens": 10307911680, "tok_s": 174270.3, "elapsed_s": 1476.3, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9860, "epoch": 0.662, "loss": 2.14774, "ppl": 8.566, "lr": 0.000103686, "lr_sched": "cosine", "grad_norm": 1.0863, "tokens": 10328862720, "tok_s": 173724.5, "elapsed_s": 1596.9, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9880, "epoch": 0.663, "loss": 2.16357, "ppl": 8.702, "lr": 0.000103162, "lr_sched": "cosine", "grad_norm": 1.0026, "tokens": 10349813760, "tok_s": 173956.9, "elapsed_s": 1717.3, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9900, "epoch": 0.665, "loss": 2.15655, "ppl": 8.641, "lr": 0.000102639, "lr_sched": "cosine", "grad_norm": 1.1483, "tokens": 10370764800, "tok_s": 174580.0, "elapsed_s": 1837.6, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9920, "epoch": 0.666, "loss": 2.12514, "ppl": 8.374, "lr": 0.000102117, "lr_sched": "cosine", "grad_norm": 1.1733, "tokens": 10391715840, "tok_s": 174216.5, "elapsed_s": 1957.8, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9940, "epoch": 0.667, "loss": 2.14179, "ppl": 8.515, "lr": 0.000101596, "lr_sched": "cosine", "grad_norm": 1.1407, "tokens": 10412666880, "tok_s": 174365.0, "elapsed_s": 2078.2, "vram_gb": 0.59, "ram_pct": 12.6, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9960, "epoch": 0.669, "loss": 2.16731, "ppl": 8.735, "lr": 0.000101077, "lr_sched": "cosine", "grad_norm": 1.3391, "tokens": 10433617920, "tok_s": 174560.5, "elapsed_s": 2198.5, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 9980, "epoch": 0.67, "loss": 2.19301, "ppl": 8.962, "lr": 0.000100558, "lr_sched": "cosine", "grad_norm": 1.5496, "tokens": 10454568960, "tok_s": 174624.7, "elapsed_s": 2318.4, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10000, "epoch": 0.671, "loss": 2.26131, "ppl": 9.596, "lr": 0.000100041, "lr_sched": "cosine", "grad_norm": 1.6754, "tokens": 10475520000, "tok_s": 173669.0, "elapsed_s": 2438.4, "vram_gb": 0.59, "ram_pct": 12.5, "disk_free_gb": 20.94, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10000, "epoch": 0.671, "val_loss": 2.32593, "val_ppl": 10.236, "best_val_ppl": 10.236, "is_best": true}
{"step": 10020, "epoch": 0.673, "loss": 2.25776, "ppl": 9.562, "lr": 9.9525e-05, "lr_sched": "cosine", "grad_norm": 1.502, "tokens": 10496471040, "tok_s": 172280.2, "elapsed_s": 3009.0, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10040, "epoch": 0.674, "loss": 2.17228, "ppl": 8.778, "lr": 9.9011e-05, "lr_sched": "cosine", "grad_norm": 1.8212, "tokens": 10517422080, "tok_s": 175035.1, "elapsed_s": 3129.2, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10060, "epoch": 0.675, "loss": 2.21225, "ppl": 9.136, "lr": 9.8497e-05, "lr_sched": "cosine", "grad_norm": 1.5756, "tokens": 10538373120, "tok_s": 173554.1, "elapsed_s": 3249.2, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10080, "epoch": 0.677, "loss": 2.21088, "ppl": 9.124, "lr": 9.7985e-05, "lr_sched": "cosine", "grad_norm": 1.2524, "tokens": 10559324160, "tok_s": 173446.2, "elapsed_s": 3369.8, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10100, "epoch": 0.678, "loss": 2.23045, "ppl": 9.304, "lr": 9.7474e-05, "lr_sched": "cosine", "grad_norm": 1.1426, "tokens": 10580275200, "tok_s": 173659.6, "elapsed_s": 3490.6, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10120, "epoch": 0.679, "loss": 2.13476, "ppl": 8.455, "lr": 9.6965e-05, "lr_sched": "cosine", "grad_norm": 1.3558, "tokens": 10601226240, "tok_s": 173874.6, "elapsed_s": 3611.0, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10140, "epoch": 0.681, "loss": 2.14651, "ppl": 8.555, "lr": 9.6457e-05, "lr_sched": "cosine", "grad_norm": 1.2798, "tokens": 10622177280, "tok_s": 172536.8, "elapsed_s": 3731.9, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10160, "epoch": 0.682, "loss": 2.19066, "ppl": 8.941, "lr": 9.595e-05, "lr_sched": "cosine", "grad_norm": 1.798, "tokens": 10643128320, "tok_s": 174044.6, "elapsed_s": 3852.9, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10180, "epoch": 0.683, "loss": 2.23534, "ppl": 9.35, "lr": 9.5444e-05, "lr_sched": "cosine", "grad_norm": 1.6453, "tokens": 10664079360, "tok_s": 174540.0, "elapsed_s": 3972.7, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10200, "epoch": 0.685, "loss": 2.189, "ppl": 8.926, "lr": 9.494e-05, "lr_sched": "cosine", "grad_norm": 1.1902, "tokens": 10685030400, "tok_s": 174768.6, "elapsed_s": 4092.7, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10220, "epoch": 0.686, "loss": 1.99675, "ppl": 7.365, "lr": 9.4437e-05, "lr_sched": "cosine", "grad_norm": 1.7964, "tokens": 10705981440, "tok_s": 174023.5, "elapsed_s": 4212.7, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10240, "epoch": 0.687, "loss": 2.20968, "ppl": 9.113, "lr": 9.3935e-05, "lr_sched": "cosine", "grad_norm": 1.1348, "tokens": 10726932480, "tok_s": 174503.2, "elapsed_s": 4332.8, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10260, "epoch": 0.689, "loss": 2.16121, "ppl": 8.682, "lr": 9.3435e-05, "lr_sched": "cosine", "grad_norm": 0.92, "tokens": 10747883520, "tok_s": 173405.1, "elapsed_s": 4453.0, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10280, "epoch": 0.69, "loss": 2.13807, "ppl": 8.483, "lr": 9.2936e-05, "lr_sched": "cosine", "grad_norm": 2.0347, "tokens": 10768834560, "tok_s": 173818.1, "elapsed_s": 4573.5, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10300, "epoch": 0.691, "loss": 2.09727, "ppl": 8.144, "lr": 9.2439e-05, "lr_sched": "cosine", "grad_norm": 1.6212, "tokens": 10789785600, "tok_s": 173876.9, "elapsed_s": 4694.1, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10320, "epoch": 0.693, "loss": 2.11778, "ppl": 8.313, "lr": 9.1943e-05, "lr_sched": "cosine", "grad_norm": 1.9219, "tokens": 10810736640, "tok_s": 174519.9, "elapsed_s": 4814.5, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10340, "epoch": 0.694, "loss": 2.13678, "ppl": 8.472, "lr": 9.1448e-05, "lr_sched": "cosine", "grad_norm": 1.7979, "tokens": 10831687680, "tok_s": 174797.3, "elapsed_s": 4934.3, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10360, "epoch": 0.695, "loss": 2.10598, "ppl": 8.215, "lr": 9.0954e-05, "lr_sched": "cosine", "grad_norm": 1.5333, "tokens": 10852638720, "tok_s": 174487.4, "elapsed_s": 5054.7, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10380, "epoch": 0.697, "loss": 2.20012, "ppl": 9.026, "lr": 9.0463e-05, "lr_sched": "cosine", "grad_norm": 1.2325, "tokens": 10873589760, "tok_s": 174562.3, "elapsed_s": 5174.8, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10400, "epoch": 0.698, "loss": 1.6673, "ppl": 5.298, "lr": 8.9972e-05, "lr_sched": "cosine", "grad_norm": 1.1839, "tokens": 10894540800, "tok_s": 174303.0, "elapsed_s": 5294.9, "vram_gb": 0.59, "ram_pct": 13.7, "disk_free_gb": 20.84, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10400, "epoch": 0.698, "val_loss": 2.50109, "val_ppl": 12.196, "best_val_ppl": 10.236, "is_best": false}
{"step": 10420, "epoch": 0.699, "loss": 1.51089, "ppl": 4.531, "lr": 8.9483e-05, "lr_sched": "cosine", "grad_norm": 1.0326, "tokens": 10915491840, "tok_s": 173447.2, "elapsed_s": 5848.1, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10440, "epoch": 0.701, "loss": 1.42532, "ppl": 4.159, "lr": 8.8995e-05, "lr_sched": "cosine", "grad_norm": 0.9875, "tokens": 10936442880, "tok_s": 173949.1, "elapsed_s": 5968.4, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10460, "epoch": 0.702, "loss": 1.37478, "ppl": 3.954, "lr": 8.8509e-05, "lr_sched": "cosine", "grad_norm": 1.2544, "tokens": 10957393920, "tok_s": 174017.9, "elapsed_s": 6088.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10480, "epoch": 0.703, "loss": 1.44645, "ppl": 4.248, "lr": 8.8024e-05, "lr_sched": "cosine", "grad_norm": 1.4754, "tokens": 10978344960, "tok_s": 175173.6, "elapsed_s": 6208.5, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10500, "epoch": 0.705, "loss": 1.31796, "ppl": 3.736, "lr": 8.7541e-05, "lr_sched": "cosine", "grad_norm": 1.0812, "tokens": 10999296000, "tok_s": 175067.5, "elapsed_s": 6327.8, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10520, "epoch": 0.706, "loss": 1.30004, "ppl": 3.669, "lr": 8.7059e-05, "lr_sched": "cosine", "grad_norm": 1.311, "tokens": 11020247040, "tok_s": 174940.0, "elapsed_s": 6447.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10540, "epoch": 0.708, "loss": 1.28571, "ppl": 3.617, "lr": 8.6579e-05, "lr_sched": "cosine", "grad_norm": 1.0005, "tokens": 11041198080, "tok_s": 174831.6, "elapsed_s": 6567.6, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10560, "epoch": 0.709, "loss": 1.27508, "ppl": 3.579, "lr": 8.61e-05, "lr_sched": "cosine", "grad_norm": 0.797, "tokens": 11062149120, "tok_s": 174153.5, "elapsed_s": 6687.6, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10580, "epoch": 0.71, "loss": 1.25271, "ppl": 3.5, "lr": 8.5623e-05, "lr_sched": "cosine", "grad_norm": 0.9525, "tokens": 11083100160, "tok_s": 174509.9, "elapsed_s": 6807.6, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10600, "epoch": 0.712, "loss": 1.24648, "ppl": 3.478, "lr": 8.5147e-05, "lr_sched": "cosine", "grad_norm": 1.4947, "tokens": 11104051200, "tok_s": 175081.7, "elapsed_s": 6927.4, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10620, "epoch": 0.713, "loss": 1.26108, "ppl": 3.529, "lr": 8.4673e-05, "lr_sched": "cosine", "grad_norm": 1.2843, "tokens": 11125002240, "tok_s": 175008.7, "elapsed_s": 7047.3, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10640, "epoch": 0.714, "loss": 1.22178, "ppl": 3.393, "lr": 8.42e-05, "lr_sched": "cosine", "grad_norm": 0.7906, "tokens": 11145953280, "tok_s": 175257.8, "elapsed_s": 7166.8, "vram_gb": 0.59, "ram_pct": 13.8, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10660, "epoch": 0.716, "loss": 1.22323, "ppl": 3.398, "lr": 8.3729e-05, "lr_sched": "cosine", "grad_norm": 0.9581, "tokens": 11166904320, "tok_s": 175028.3, "elapsed_s": 7286.6, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10680, "epoch": 0.717, "loss": 1.20206, "ppl": 3.327, "lr": 8.3259e-05, "lr_sched": "cosine", "grad_norm": 0.8588, "tokens": 11187855360, "tok_s": 175406.6, "elapsed_s": 7406.2, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10700, "epoch": 0.718, "loss": 1.22942, "ppl": 3.419, "lr": 8.2791e-05, "lr_sched": "cosine", "grad_norm": 0.9866, "tokens": 11208806400, "tok_s": 175096.7, "elapsed_s": 7525.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10720, "epoch": 0.72, "loss": 1.18375, "ppl": 3.267, "lr": 8.2324e-05, "lr_sched": "cosine", "grad_norm": 1.0252, "tokens": 11229757440, "tok_s": 174672.7, "elapsed_s": 7645.6, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10740, "epoch": 0.721, "loss": 1.21195, "ppl": 3.36, "lr": 8.1859e-05, "lr_sched": "cosine", "grad_norm": 1.178, "tokens": 11250708480, "tok_s": 174184.9, "elapsed_s": 7765.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10760, "epoch": 0.722, "loss": 1.17846, "ppl": 3.249, "lr": 8.1396e-05, "lr_sched": "cosine", "grad_norm": 1.0654, "tokens": 11271659520, "tok_s": 173852.5, "elapsed_s": 7885.9, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10780, "epoch": 0.724, "loss": 1.19427, "ppl": 3.301, "lr": 8.0934e-05, "lr_sched": "cosine", "grad_norm": 1.0703, "tokens": 11292610560, "tok_s": 173525.5, "elapsed_s": 8006.5, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10800, "epoch": 0.725, "loss": 1.19338, "ppl": 3.298, "lr": 8.0474e-05, "lr_sched": "cosine", "grad_norm": 1.427, "tokens": 11313561600, "tok_s": 174656.1, "elapsed_s": 8126.3, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10800, "epoch": 0.725, "val_loss": 2.73388, "val_ppl": 15.392, "best_val_ppl": 10.236, "is_best": false}
{"step": 10820, "epoch": 0.726, "loss": 1.22911, "ppl": 3.418, "lr": 8.0015e-05, "lr_sched": "cosine", "grad_norm": 1.636, "tokens": 11334512640, "tok_s": 175848.7, "elapsed_s": 8677.6, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10840, "epoch": 0.728, "loss": 1.16075, "ppl": 3.192, "lr": 7.9558e-05, "lr_sched": "cosine", "grad_norm": 1.0744, "tokens": 11355463680, "tok_s": 175376.9, "elapsed_s": 8796.9, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10860, "epoch": 0.729, "loss": 1.21122, "ppl": 3.358, "lr": 7.9103e-05, "lr_sched": "cosine", "grad_norm": 0.9433, "tokens": 11376414720, "tok_s": 174997.8, "elapsed_s": 8916.4, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10880, "epoch": 0.73, "loss": 1.18242, "ppl": 3.262, "lr": 7.8649e-05, "lr_sched": "cosine", "grad_norm": 0.885, "tokens": 11397365760, "tok_s": 175113.9, "elapsed_s": 9036.0, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10900, "epoch": 0.732, "loss": 1.16676, "ppl": 3.212, "lr": 7.8197e-05, "lr_sched": "cosine", "grad_norm": 1.1557, "tokens": 11418316800, "tok_s": 174888.6, "elapsed_s": 9155.3, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10920, "epoch": 0.733, "loss": 1.16374, "ppl": 3.202, "lr": 7.7747e-05, "lr_sched": "cosine", "grad_norm": 0.9415, "tokens": 11439267840, "tok_s": 174995.5, "elapsed_s": 9274.9, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10940, "epoch": 0.734, "loss": 1.14084, "ppl": 3.129, "lr": 7.7298e-05, "lr_sched": "cosine", "grad_norm": 1.1499, "tokens": 11460218880, "tok_s": 174743.8, "elapsed_s": 9394.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10960, "epoch": 0.736, "loss": 1.15103, "ppl": 3.161, "lr": 7.6851e-05, "lr_sched": "cosine", "grad_norm": 1.0301, "tokens": 11481169920, "tok_s": 175184.6, "elapsed_s": 9514.5, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 10980, "epoch": 0.737, "loss": 1.13872, "ppl": 3.123, "lr": 7.6406e-05, "lr_sched": "cosine", "grad_norm": 0.7846, "tokens": 11502120960, "tok_s": 175072.7, "elapsed_s": 9634.1, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11000, "epoch": 0.738, "loss": 1.14554, "ppl": 3.144, "lr": 7.5962e-05, "lr_sched": "cosine", "grad_norm": 0.8048, "tokens": 11523072000, "tok_s": 175148.8, "elapsed_s": 9753.6, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11020, "epoch": 0.74, "loss": 1.12023, "ppl": 3.066, "lr": 7.552e-05, "lr_sched": "cosine", "grad_norm": 0.9806, "tokens": 11544023040, "tok_s": 175345.8, "elapsed_s": 9872.9, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11040, "epoch": 0.741, "loss": 1.09239, "ppl": 2.981, "lr": 7.508e-05, "lr_sched": "cosine", "grad_norm": 1.0576, "tokens": 11564974080, "tok_s": 175593.8, "elapsed_s": 9992.3, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11060, "epoch": 0.742, "loss": 1.10462, "ppl": 3.018, "lr": 7.4641e-05, "lr_sched": "cosine", "grad_norm": 0.8206, "tokens": 11585925120, "tok_s": 175583.2, "elapsed_s": 10111.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11080, "epoch": 0.744, "loss": 1.10772, "ppl": 3.027, "lr": 7.4204e-05, "lr_sched": "cosine", "grad_norm": 0.8161, "tokens": 11606876160, "tok_s": 174620.7, "elapsed_s": 10231.2, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11100, "epoch": 0.745, "loss": 1.13645, "ppl": 3.116, "lr": 7.3769e-05, "lr_sched": "cosine", "grad_norm": 1.047, "tokens": 11627827200, "tok_s": 175986.7, "elapsed_s": 10350.9, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11120, "epoch": 0.746, "loss": 1.11829, "ppl": 3.06, "lr": 7.3336e-05, "lr_sched": "cosine", "grad_norm": 0.8549, "tokens": 11648778240, "tok_s": 176491.2, "elapsed_s": 10470.3, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11140, "epoch": 0.748, "loss": 1.12099, "ppl": 3.068, "lr": 7.2904e-05, "lr_sched": "cosine", "grad_norm": 0.9089, "tokens": 11669729280, "tok_s": 175790.2, "elapsed_s": 10589.5, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11160, "epoch": 0.749, "loss": 1.21119, "ppl": 3.357, "lr": 7.2474e-05, "lr_sched": "cosine", "grad_norm": 1.1333, "tokens": 11690680320, "tok_s": 174346.3, "elapsed_s": 10709.3, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11180, "epoch": 0.75, "loss": 2.50691, "ppl": 12.267, "lr": 7.2046e-05, "lr_sched": "cosine", "grad_norm": 2.0311, "tokens": 11711631360, "tok_s": 175201.0, "elapsed_s": 10829.1, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11200, "epoch": 0.752, "loss": 2.3412, "ppl": 10.394, "lr": 7.162e-05, "lr_sched": "cosine", "grad_norm": 1.3261, "tokens": 11732582400, "tok_s": 174902.2, "elapsed_s": 10948.8, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11200, "epoch": 0.752, "val_loss": 2.59169, "val_ppl": 13.352, "best_val_ppl": 10.236, "is_best": false}
{"step": 11220, "epoch": 0.753, "loss": 2.26152, "ppl": 9.598, "lr": 7.1195e-05, "lr_sched": "cosine", "grad_norm": 1.2509, "tokens": 11753533440, "tok_s": 174061.6, "elapsed_s": 11502.8, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11240, "epoch": 0.755, "loss": 2.29601, "ppl": 9.934, "lr": 7.0772e-05, "lr_sched": "cosine", "grad_norm": 1.3778, "tokens": 11774484480, "tok_s": 172772.6, "elapsed_s": 11623.0, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11260, "epoch": 0.756, "loss": 2.30014, "ppl": 9.976, "lr": 7.0351e-05, "lr_sched": "cosine", "grad_norm": 1.3422, "tokens": 11795435520, "tok_s": 174859.1, "elapsed_s": 11743.4, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11280, "epoch": 0.757, "loss": 2.28538, "ppl": 9.829, "lr": 6.9932e-05, "lr_sched": "cosine", "grad_norm": 1.2655, "tokens": 11816386560, "tok_s": 174258.3, "elapsed_s": 11863.1, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11300, "epoch": 0.759, "loss": 2.26099, "ppl": 9.593, "lr": 6.9515e-05, "lr_sched": "cosine", "grad_norm": 1.454, "tokens": 11837337600, "tok_s": 173380.9, "elapsed_s": 11983.7, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11320, "epoch": 0.76, "loss": 2.20027, "ppl": 9.027, "lr": 6.9099e-05, "lr_sched": "cosine", "grad_norm": 1.492, "tokens": 11858288640, "tok_s": 175266.3, "elapsed_s": 12103.8, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11340, "epoch": 0.761, "loss": 2.23721, "ppl": 9.367, "lr": 6.8685e-05, "lr_sched": "cosine", "grad_norm": 0.9066, "tokens": 11879239680, "tok_s": 175269.6, "elapsed_s": 12223.4, "vram_gb": 0.59, "ram_pct": 13.9, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11360, "epoch": 0.763, "loss": 2.17661, "ppl": 8.816, "lr": 6.8273e-05, "lr_sched": "cosine", "grad_norm": 2.0059, "tokens": 11900190720, "tok_s": 172314.3, "elapsed_s": 12343.6, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11380, "epoch": 0.764, "loss": 1.87538, "ppl": 6.523, "lr": 6.7863e-05, "lr_sched": "cosine", "grad_norm": 1.3869, "tokens": 11921141760, "tok_s": 175186.2, "elapsed_s": 12463.8, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11400, "epoch": 0.765, "loss": 2.18644, "ppl": 8.903, "lr": 6.7455e-05, "lr_sched": "cosine", "grad_norm": 1.1085, "tokens": 11942092800, "tok_s": 175426.7, "elapsed_s": 12583.2, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11420, "epoch": 0.767, "loss": 2.14675, "ppl": 8.557, "lr": 6.7049e-05, "lr_sched": "cosine", "grad_norm": 1.1251, "tokens": 11963043840, "tok_s": 173951.8, "elapsed_s": 12703.1, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11440, "epoch": 0.768, "loss": 2.17469, "ppl": 8.799, "lr": 6.6644e-05, "lr_sched": "cosine", "grad_norm": 1.0916, "tokens": 11983994880, "tok_s": 175478.6, "elapsed_s": 12823.1, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11460, "epoch": 0.769, "loss": 2.22706, "ppl": 9.273, "lr": 6.6241e-05, "lr_sched": "cosine", "grad_norm": 1.5164, "tokens": 12004945920, "tok_s": 175017.1, "elapsed_s": 12942.7, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11480, "epoch": 0.771, "loss": 2.18128, "ppl": 8.858, "lr": 6.5841e-05, "lr_sched": "cosine", "grad_norm": 1.7166, "tokens": 12025896960, "tok_s": 174467.0, "elapsed_s": 13062.8, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11500, "epoch": 0.772, "loss": 2.2304, "ppl": 9.304, "lr": 6.5442e-05, "lr_sched": "cosine", "grad_norm": 1.57, "tokens": 12046848000, "tok_s": 174551.7, "elapsed_s": 13183.3, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11520, "epoch": 0.773, "loss": 2.19562, "ppl": 8.986, "lr": 6.5045e-05, "lr_sched": "cosine", "grad_norm": 1.6708, "tokens": 12067799040, "tok_s": 174715.9, "elapsed_s": 13303.0, "vram_gb": 0.59, "ram_pct": 14.1, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11540, "epoch": 0.775, "loss": 2.17808, "ppl": 8.829, "lr": 6.465e-05, "lr_sched": "cosine", "grad_norm": 1.0769, "tokens": 12088750080, "tok_s": 173842.0, "elapsed_s": 13422.9, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11560, "epoch": 0.776, "loss": 2.44553, "ppl": 11.537, "lr": 6.4257e-05, "lr_sched": "cosine", "grad_norm": 1.5102, "tokens": 12109701120, "tok_s": 174693.3, "elapsed_s": 13543.4, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11580, "epoch": 0.777, "loss": 2.34586, "ppl": 10.442, "lr": 6.3865e-05, "lr_sched": "cosine", "grad_norm": 0.8121, "tokens": 12130652160, "tok_s": 175394.7, "elapsed_s": 13663.1, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11600, "epoch": 0.779, "loss": 2.2836, "ppl": 9.812, "lr": 6.3476e-05, "lr_sched": "cosine", "grad_norm": 0.9228, "tokens": 12151603200, "tok_s": 174087.8, "elapsed_s": 13782.9, "vram_gb": 0.59, "ram_pct": 14.0, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11600, "epoch": 0.779, "val_loss": 2.34164, "val_ppl": 10.398, "best_val_ppl": 10.236, "is_best": false}
{"step": 11620, "epoch": 0.78, "loss": 2.32963, "ppl": 10.274, "lr": 6.3089e-05, "lr_sched": "cosine", "grad_norm": 1.2663, "tokens": 12172554240, "tok_s": 174551.8, "elapsed_s": 14336.0, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11640, "epoch": 0.781, "loss": 2.29229, "ppl": 9.898, "lr": 6.2703e-05, "lr_sched": "cosine", "grad_norm": 0.8639, "tokens": 12193505280, "tok_s": 173093.6, "elapsed_s": 14456.5, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11660, "epoch": 0.783, "loss": 2.25514, "ppl": 9.537, "lr": 6.232e-05, "lr_sched": "cosine", "grad_norm": 1.6307, "tokens": 12214456320, "tok_s": 175785.7, "elapsed_s": 14576.4, "vram_gb": 0.59, "ram_pct": 14.5, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11680, "epoch": 0.784, "loss": 2.34632, "ppl": 10.447, "lr": 6.1938e-05, "lr_sched": "cosine", "grad_norm": 1.0266, "tokens": 12235407360, "tok_s": 173409.8, "elapsed_s": 14696.9, "vram_gb": 0.59, "ram_pct": 14.5, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11700, "epoch": 0.785, "loss": 2.3212, "ppl": 10.188, "lr": 6.1559e-05, "lr_sched": "cosine", "grad_norm": 0.9271, "tokens": 12256358400, "tok_s": 171723.0, "elapsed_s": 14817.9, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11720, "epoch": 0.787, "loss": 2.22265, "ppl": 9.232, "lr": 6.1181e-05, "lr_sched": "cosine", "grad_norm": 0.8959, "tokens": 12277309440, "tok_s": 174860.9, "elapsed_s": 14938.0, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11740, "epoch": 0.788, "loss": 2.31246, "ppl": 10.099, "lr": 6.0806e-05, "lr_sched": "cosine", "grad_norm": 1.1541, "tokens": 12298260480, "tok_s": 174842.6, "elapsed_s": 15057.7, "vram_gb": 0.59, "ram_pct": 14.5, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11760, "epoch": 0.789, "loss": 2.30828, "ppl": 10.057, "lr": 6.0432e-05, "lr_sched": "cosine", "grad_norm": 1.0826, "tokens": 12319211520, "tok_s": 174312.5, "elapsed_s": 15178.5, "vram_gb": 0.59, "ram_pct": 14.5, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11780, "epoch": 0.791, "loss": 2.30949, "ppl": 10.069, "lr": 6.006e-05, "lr_sched": "cosine", "grad_norm": 1.2313, "tokens": 12340162560, "tok_s": 175139.9, "elapsed_s": 15298.2, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11800, "epoch": 0.792, "loss": 2.31365, "ppl": 10.111, "lr": 5.9691e-05, "lr_sched": "cosine", "grad_norm": 0.8331, "tokens": 12361113600, "tok_s": 173844.1, "elapsed_s": 15419.0, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11820, "epoch": 0.793, "loss": 2.2965, "ppl": 9.939, "lr": 5.9323e-05, "lr_sched": "cosine", "grad_norm": 0.9486, "tokens": 12382064640, "tok_s": 174431.0, "elapsed_s": 15539.3, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11840, "epoch": 0.795, "loss": 2.26984, "ppl": 9.678, "lr": 5.8957e-05, "lr_sched": "cosine", "grad_norm": 0.9943, "tokens": 12403015680, "tok_s": 174421.9, "elapsed_s": 15659.3, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11860, "epoch": 0.796, "loss": 2.24327, "ppl": 9.424, "lr": 5.8594e-05, "lr_sched": "cosine", "grad_norm": 0.9107, "tokens": 12423966720, "tok_s": 175056.9, "elapsed_s": 15779.2, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11880, "epoch": 0.797, "loss": 2.28648, "ppl": 9.84, "lr": 5.8232e-05, "lr_sched": "cosine", "grad_norm": 1.1743, "tokens": 12444917760, "tok_s": 175160.2, "elapsed_s": 15898.9, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11900, "epoch": 0.799, "loss": 2.27878, "ppl": 9.765, "lr": 5.7873e-05, "lr_sched": "cosine", "grad_norm": 1.0467, "tokens": 12465868800, "tok_s": 174758.8, "elapsed_s": 16018.7, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11920, "epoch": 0.8, "loss": 2.23233, "ppl": 9.322, "lr": 5.7515e-05, "lr_sched": "cosine", "grad_norm": 1.2794, "tokens": 12486819840, "tok_s": 173048.3, "elapsed_s": 16138.8, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11940, "epoch": 0.802, "loss": 2.30188, "ppl": 9.993, "lr": 5.716e-05, "lr_sched": "cosine", "grad_norm": 1.1625, "tokens": 12507770880, "tok_s": 175051.1, "elapsed_s": 16259.4, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11960, "epoch": 0.803, "loss": 2.26962, "ppl": 9.676, "lr": 5.6806e-05, "lr_sched": "cosine", "grad_norm": 0.8839, "tokens": 12528721920, "tok_s": 173150.7, "elapsed_s": 16379.8, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 11980, "epoch": 0.804, "loss": 2.22159, "ppl": 9.222, "lr": 5.6455e-05, "lr_sched": "cosine", "grad_norm": 1.2691, "tokens": 12549672960, "tok_s": 174933.8, "elapsed_s": 16499.9, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12000, "epoch": 0.806, "loss": 2.28679, "ppl": 9.843, "lr": 5.6106e-05, "lr_sched": "cosine", "grad_norm": 0.8506, "tokens": 12570624000, "tok_s": 175004.4, "elapsed_s": 16619.7, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12000, "epoch": 0.806, "val_loss": 2.27037, "val_ppl": 9.683, "best_val_ppl": 9.683, "is_best": true}
{"step": 12020, "epoch": 0.807, "loss": 2.27161, "ppl": 9.695, "lr": 5.5758e-05, "lr_sched": "cosine", "grad_norm": 0.9502, "tokens": 12591575040, "tok_s": 173389.5, "elapsed_s": 17172.2, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12040, "epoch": 0.808, "loss": 2.22002, "ppl": 9.207, "lr": 5.5413e-05, "lr_sched": "cosine", "grad_norm": 0.8955, "tokens": 12612526080, "tok_s": 172606.6, "elapsed_s": 17292.1, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12060, "epoch": 0.81, "loss": 2.30287, "ppl": 10.003, "lr": 5.507e-05, "lr_sched": "cosine", "grad_norm": 1.436, "tokens": 12633477120, "tok_s": 174101.3, "elapsed_s": 17413.0, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12080, "epoch": 0.811, "loss": 2.25488, "ppl": 9.534, "lr": 5.4729e-05, "lr_sched": "cosine", "grad_norm": 1.0826, "tokens": 12654428160, "tok_s": 175724.7, "elapsed_s": 17532.8, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12100, "epoch": 0.812, "loss": 2.20791, "ppl": 9.097, "lr": 5.4391e-05, "lr_sched": "cosine", "grad_norm": 1.2522, "tokens": 12675379200, "tok_s": 172586.6, "elapsed_s": 17653.0, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12120, "epoch": 0.814, "loss": 2.2815, "ppl": 9.791, "lr": 5.4054e-05, "lr_sched": "cosine", "grad_norm": 1.1983, "tokens": 12696330240, "tok_s": 174426.4, "elapsed_s": 17773.5, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12140, "epoch": 0.815, "loss": 2.24461, "ppl": 9.437, "lr": 5.3719e-05, "lr_sched": "cosine", "grad_norm": 1.0332, "tokens": 12717281280, "tok_s": 174345.4, "elapsed_s": 17893.5, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12160, "epoch": 0.816, "loss": 2.21752, "ppl": 9.184, "lr": 5.3387e-05, "lr_sched": "cosine", "grad_norm": 1.2412, "tokens": 12738232320, "tok_s": 174303.9, "elapsed_s": 18013.2, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12180, "epoch": 0.818, "loss": 2.2873, "ppl": 9.848, "lr": 5.3056e-05, "lr_sched": "cosine", "grad_norm": 1.5955, "tokens": 12759183360, "tok_s": 173887.4, "elapsed_s": 18133.2, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12200, "epoch": 0.819, "loss": 2.26585, "ppl": 9.639, "lr": 5.2728e-05, "lr_sched": "cosine", "grad_norm": 0.8256, "tokens": 12780134400, "tok_s": 173501.2, "elapsed_s": 18254.5, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12220, "epoch": 0.82, "loss": 2.25184, "ppl": 9.505, "lr": 5.2402e-05, "lr_sched": "cosine", "grad_norm": 0.9993, "tokens": 12801085440, "tok_s": 175147.5, "elapsed_s": 18374.8, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12240, "epoch": 0.822, "loss": 2.2721, "ppl": 9.7, "lr": 5.2078e-05, "lr_sched": "cosine", "grad_norm": 1.0502, "tokens": 12822036480, "tok_s": 174821.8, "elapsed_s": 18494.6, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12260, "epoch": 0.823, "loss": 2.27274, "ppl": 9.706, "lr": 5.1756e-05, "lr_sched": "cosine", "grad_norm": 0.9567, "tokens": 12842987520, "tok_s": 172167.1, "elapsed_s": 18615.3, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12280, "epoch": 0.824, "loss": 2.25606, "ppl": 9.545, "lr": 5.1437e-05, "lr_sched": "cosine", "grad_norm": 0.8796, "tokens": 12863938560, "tok_s": 174724.1, "elapsed_s": 18736.5, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12300, "epoch": 0.826, "loss": 2.20398, "ppl": 9.061, "lr": 5.1119e-05, "lr_sched": "cosine", "grad_norm": 1.5662, "tokens": 12884889600, "tok_s": 174592.5, "elapsed_s": 18856.3, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12320, "epoch": 0.827, "loss": 2.27602, "ppl": 9.738, "lr": 5.0804e-05, "lr_sched": "cosine", "grad_norm": 0.9774, "tokens": 12905840640, "tok_s": 172452.3, "elapsed_s": 18977.4, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12340, "epoch": 0.828, "loss": 2.25588, "ppl": 9.544, "lr": 5.0491e-05, "lr_sched": "cosine", "grad_norm": 0.8656, "tokens": 12926791680, "tok_s": 174869.8, "elapsed_s": 19097.6, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12360, "epoch": 0.83, "loss": 2.22659, "ppl": 9.268, "lr": 5.018e-05, "lr_sched": "cosine", "grad_norm": 0.9551, "tokens": 12947742720, "tok_s": 174880.0, "elapsed_s": 19217.5, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12380, "epoch": 0.831, "loss": 2.32209, "ppl": 10.197, "lr": 4.9871e-05, "lr_sched": "cosine", "grad_norm": 1.0969, "tokens": 12968693760, "tok_s": 174893.8, "elapsed_s": 19337.3, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12400, "epoch": 0.832, "loss": 2.23109, "ppl": 9.31, "lr": 4.9564e-05, "lr_sched": "cosine", "grad_norm": 1.1255, "tokens": 12989644800, "tok_s": 174933.6, "elapsed_s": 19457.1, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.81, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12400, "epoch": 0.832, "val_loss": 2.25552, "val_ppl": 9.54, "best_val_ppl": 9.54, "is_best": true}
{"step": 12420, "epoch": 0.834, "loss": 2.26318, "ppl": 9.614, "lr": 4.926e-05, "lr_sched": "cosine", "grad_norm": 0.8803, "tokens": 13010595840, "tok_s": 172817.0, "elapsed_s": 20011.2, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12440, "epoch": 0.835, "loss": 2.18276, "ppl": 8.871, "lr": 4.8958e-05, "lr_sched": "cosine", "grad_norm": 1.0878, "tokens": 13031546880, "tok_s": 175061.8, "elapsed_s": 20131.4, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12460, "epoch": 0.836, "loss": 2.25661, "ppl": 9.551, "lr": 4.8658e-05, "lr_sched": "cosine", "grad_norm": 1.1435, "tokens": 13052497920, "tok_s": 175208.3, "elapsed_s": 20251.2, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12480, "epoch": 0.838, "loss": 2.24917, "ppl": 9.48, "lr": 4.836e-05, "lr_sched": "cosine", "grad_norm": 1.2451, "tokens": 13073448960, "tok_s": 174916.6, "elapsed_s": 20370.9, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12500, "epoch": 0.839, "loss": 2.19362, "ppl": 8.968, "lr": 4.8064e-05, "lr_sched": "cosine", "grad_norm": 0.9001, "tokens": 13094400000, "tok_s": 173109.1, "elapsed_s": 20491.1, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12520, "epoch": 0.84, "loss": 2.33395, "ppl": 10.319, "lr": 4.7771e-05, "lr_sched": "cosine", "grad_norm": 1.0433, "tokens": 13115351040, "tok_s": 174670.1, "elapsed_s": 20612.1, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12540, "epoch": 0.842, "loss": 2.26155, "ppl": 9.598, "lr": 4.748e-05, "lr_sched": "cosine", "grad_norm": 1.1708, "tokens": 13136302080, "tok_s": 174992.6, "elapsed_s": 20731.9, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12560, "epoch": 0.843, "loss": 2.22743, "ppl": 9.276, "lr": 4.7191e-05, "lr_sched": "cosine", "grad_norm": 1.0571, "tokens": 13157253120, "tok_s": 173057.5, "elapsed_s": 20852.1, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12580, "epoch": 0.844, "loss": 2.25593, "ppl": 9.544, "lr": 4.6905e-05, "lr_sched": "cosine", "grad_norm": 0.9769, "tokens": 13178204160, "tok_s": 175082.7, "elapsed_s": 20972.6, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12600, "epoch": 0.846, "loss": 2.29733, "ppl": 9.948, "lr": 4.662e-05, "lr_sched": "cosine", "grad_norm": 1.1601, "tokens": 13199155200, "tok_s": 174688.4, "elapsed_s": 21092.4, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12620, "epoch": 0.847, "loss": 2.24284, "ppl": 9.42, "lr": 4.6338e-05, "lr_sched": "cosine", "grad_norm": 1.1757, "tokens": 13220106240, "tok_s": 173493.7, "elapsed_s": 21213.6, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12640, "epoch": 0.848, "loss": 2.26238, "ppl": 9.606, "lr": 4.6058e-05, "lr_sched": "cosine", "grad_norm": 0.8388, "tokens": 13241057280, "tok_s": 174754.4, "elapsed_s": 21333.6, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12660, "epoch": 0.85, "loss": 2.28085, "ppl": 9.785, "lr": 4.5781e-05, "lr_sched": "cosine", "grad_norm": 0.8, "tokens": 13262008320, "tok_s": 174577.5, "elapsed_s": 21453.6, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12680, "epoch": 0.851, "loss": 2.22668, "ppl": 9.269, "lr": 4.5506e-05, "lr_sched": "cosine", "grad_norm": 1.1778, "tokens": 13282959360, "tok_s": 174806.9, "elapsed_s": 21573.5, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12700, "epoch": 0.853, "loss": 2.26315, "ppl": 9.613, "lr": 4.5233e-05, "lr_sched": "cosine", "grad_norm": 0.9437, "tokens": 13303910400, "tok_s": 174788.1, "elapsed_s": 21693.4, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12720, "epoch": 0.854, "loss": 2.26156, "ppl": 9.598, "lr": 4.4962e-05, "lr_sched": "cosine", "grad_norm": 1.0481, "tokens": 13324861440, "tok_s": 174621.3, "elapsed_s": 21813.2, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12740, "epoch": 0.855, "loss": 2.27654, "ppl": 9.743, "lr": 4.4693e-05, "lr_sched": "cosine", "grad_norm": 0.8535, "tokens": 13345812480, "tok_s": 174240.8, "elapsed_s": 21933.1, "vram_gb": 0.59, "ram_pct": 14.6, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12760, "epoch": 0.857, "loss": 2.30416, "ppl": 10.016, "lr": 4.4427e-05, "lr_sched": "cosine", "grad_norm": 0.975, "tokens": 13366763520, "tok_s": 174707.5, "elapsed_s": 22054.0, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12780, "epoch": 0.858, "loss": 2.24652, "ppl": 9.455, "lr": 4.4163e-05, "lr_sched": "cosine", "grad_norm": 0.8752, "tokens": 13387714560, "tok_s": 174913.2, "elapsed_s": 22173.7, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12800, "epoch": 0.859, "loss": 2.26294, "ppl": 9.611, "lr": 4.3902e-05, "lr_sched": "cosine", "grad_norm": 0.9348, "tokens": 13408665600, "tok_s": 172890.6, "elapsed_s": 22293.7, "vram_gb": 0.59, "ram_pct": 14.7, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12800, "epoch": 0.859, "val_loss": 2.24795, "val_ppl": 9.468, "best_val_ppl": 9.468, "is_best": true}
{"step": 12820, "epoch": 0.861, "loss": 2.26771, "ppl": 9.657, "lr": 4.3643e-05, "lr_sched": "cosine", "grad_norm": 0.9916, "tokens": 13429616640, "tok_s": 171102.6, "elapsed_s": 22846.7, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12840, "epoch": 0.862, "loss": 2.25466, "ppl": 9.532, "lr": 4.3386e-05, "lr_sched": "cosine", "grad_norm": 1.0917, "tokens": 13450567680, "tok_s": 173172.7, "elapsed_s": 22968.3, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12860, "epoch": 0.863, "loss": 2.23601, "ppl": 9.356, "lr": 4.3131e-05, "lr_sched": "cosine", "grad_norm": 0.9768, "tokens": 13471518720, "tok_s": 174636.6, "elapsed_s": 23088.3, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12880, "epoch": 0.865, "loss": 2.30049, "ppl": 9.979, "lr": 4.2879e-05, "lr_sched": "cosine", "grad_norm": 1.0383, "tokens": 13492469760, "tok_s": 174400.7, "elapsed_s": 23208.1, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12900, "epoch": 0.866, "loss": 2.25134, "ppl": 9.5, "lr": 4.2629e-05, "lr_sched": "cosine", "grad_norm": 0.8409, "tokens": 13513420800, "tok_s": 174774.7, "elapsed_s": 23328.1, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12920, "epoch": 0.867, "loss": 2.23737, "ppl": 9.369, "lr": 4.2381e-05, "lr_sched": "cosine", "grad_norm": 0.8319, "tokens": 13534371840, "tok_s": 174907.9, "elapsed_s": 23447.9, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12940, "epoch": 0.869, "loss": 2.28016, "ppl": 9.778, "lr": 4.2136e-05, "lr_sched": "cosine", "grad_norm": 0.9948, "tokens": 13555322880, "tok_s": 174014.2, "elapsed_s": 23567.9, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12960, "epoch": 0.87, "loss": 2.29036, "ppl": 9.879, "lr": 4.1893e-05, "lr_sched": "cosine", "grad_norm": 1.0527, "tokens": 13576273920, "tok_s": 174030.3, "elapsed_s": 23688.3, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 12980, "epoch": 0.871, "loss": 2.21593, "ppl": 9.17, "lr": 4.1652e-05, "lr_sched": "cosine", "grad_norm": 0.8489, "tokens": 13597224960, "tok_s": 174256.2, "elapsed_s": 23808.7, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13000, "epoch": 0.873, "loss": 2.2685, "ppl": 9.665, "lr": 4.1414e-05, "lr_sched": "cosine", "grad_norm": 1.0317, "tokens": 13618176000, "tok_s": 173436.5, "elapsed_s": 23929.2, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13020, "epoch": 0.874, "loss": 2.28062, "ppl": 9.783, "lr": 4.1178e-05, "lr_sched": "cosine", "grad_norm": 0.9754, "tokens": 13639127040, "tok_s": 173943.3, "elapsed_s": 24049.4, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13040, "epoch": 0.875, "loss": 2.25584, "ppl": 9.543, "lr": 4.0944e-05, "lr_sched": "cosine", "grad_norm": 0.9104, "tokens": 13660078080, "tok_s": 172566.7, "elapsed_s": 24170.4, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13060, "epoch": 0.877, "loss": 2.29675, "ppl": 9.942, "lr": 4.0713e-05, "lr_sched": "cosine", "grad_norm": 0.8919, "tokens": 13681029120, "tok_s": 173191.8, "elapsed_s": 24291.7, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13080, "epoch": 0.878, "loss": 2.25811, "ppl": 9.565, "lr": 4.0484e-05, "lr_sched": "cosine", "grad_norm": 1.0081, "tokens": 13701980160, "tok_s": 173605.8, "elapsed_s": 24412.6, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13100, "epoch": 0.879, "loss": 2.23574, "ppl": 9.353, "lr": 4.0258e-05, "lr_sched": "cosine", "grad_norm": 1.0886, "tokens": 13722931200, "tok_s": 174308.5, "elapsed_s": 24533.4, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13120, "epoch": 0.881, "loss": 2.26431, "ppl": 9.624, "lr": 4.0034e-05, "lr_sched": "cosine", "grad_norm": 2.202, "tokens": 13743882240, "tok_s": 175063.1, "elapsed_s": 24653.2, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13140, "epoch": 0.882, "loss": 2.24463, "ppl": 9.437, "lr": 3.9812e-05, "lr_sched": "cosine", "grad_norm": 1.2109, "tokens": 13764833280, "tok_s": 174937.5, "elapsed_s": 24773.0, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13160, "epoch": 0.883, "loss": 2.27906, "ppl": 9.767, "lr": 3.9593e-05, "lr_sched": "cosine", "grad_norm": 1.3918, "tokens": 13785784320, "tok_s": 174907.1, "elapsed_s": 24892.7, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13180, "epoch": 0.885, "loss": 2.24748, "ppl": 9.464, "lr": 3.9376e-05, "lr_sched": "cosine", "grad_norm": 1.1402, "tokens": 13806735360, "tok_s": 173932.7, "elapsed_s": 25012.8, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13200, "epoch": 0.886, "loss": 2.17751, "ppl": 8.824, "lr": 3.9161e-05, "lr_sched": "cosine", "grad_norm": 1.6545, "tokens": 13827686400, "tok_s": 174870.0, "elapsed_s": 25132.7, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13200, "epoch": 0.886, "val_loss": 2.31707, "val_ppl": 10.146, "best_val_ppl": 9.468, "is_best": false}
{"step": 13220, "epoch": 0.887, "loss": 2.22759, "ppl": 9.277, "lr": 3.8949e-05, "lr_sched": "cosine", "grad_norm": 1.0647, "tokens": 13848637440, "tok_s": 173043.0, "elapsed_s": 25683.8, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13240, "epoch": 0.889, "loss": 2.1931, "ppl": 8.963, "lr": 3.8739e-05, "lr_sched": "cosine", "grad_norm": 1.0236, "tokens": 13869588480, "tok_s": 173929.0, "elapsed_s": 25805.1, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13260, "epoch": 0.89, "loss": 2.20185, "ppl": 9.042, "lr": 3.8532e-05, "lr_sched": "cosine", "grad_norm": 1.2465, "tokens": 13890539520, "tok_s": 174496.6, "elapsed_s": 25925.3, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13280, "epoch": 0.891, "loss": 2.1968, "ppl": 8.996, "lr": 3.8327e-05, "lr_sched": "cosine", "grad_norm": 0.9816, "tokens": 13911490560, "tok_s": 174664.3, "elapsed_s": 26045.2, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13300, "epoch": 0.893, "loss": 2.11946, "ppl": 8.327, "lr": 3.8125e-05, "lr_sched": "cosine", "grad_norm": 1.1978, "tokens": 13932441600, "tok_s": 174765.7, "elapsed_s": 26164.9, "vram_gb": 0.59, "ram_pct": 14.8, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13320, "epoch": 0.894, "loss": 2.17719, "ppl": 8.822, "lr": 3.7925e-05, "lr_sched": "cosine", "grad_norm": 1.3533, "tokens": 13953392640, "tok_s": 175071.5, "elapsed_s": 26284.7, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13340, "epoch": 0.895, "loss": 2.13955, "ppl": 8.496, "lr": 3.7727e-05, "lr_sched": "cosine", "grad_norm": 1.094, "tokens": 13974343680, "tok_s": 175004.5, "elapsed_s": 26404.5, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13360, "epoch": 0.897, "loss": 2.11539, "ppl": 8.293, "lr": 3.7532e-05, "lr_sched": "cosine", "grad_norm": 1.1606, "tokens": 13995294720, "tok_s": 175347.0, "elapsed_s": 26524.2, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13380, "epoch": 0.898, "loss": 2.16469, "ppl": 8.712, "lr": 3.7339e-05, "lr_sched": "cosine", "grad_norm": 1.1297, "tokens": 14016245760, "tok_s": 175201.1, "elapsed_s": 26643.8, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13400, "epoch": 0.9, "loss": 2.24974, "ppl": 9.485, "lr": 3.7148e-05, "lr_sched": "cosine", "grad_norm": 1.1896, "tokens": 14037196800, "tok_s": 174269.5, "elapsed_s": 26763.6, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13420, "epoch": 0.901, "loss": 2.21632, "ppl": 9.174, "lr": 3.6961e-05, "lr_sched": "cosine", "grad_norm": 1.3977, "tokens": 14058147840, "tok_s": 172743.4, "elapsed_s": 26884.8, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13440, "epoch": 0.902, "loss": 2.19597, "ppl": 8.989, "lr": 3.6775e-05, "lr_sched": "cosine", "grad_norm": 1.7845, "tokens": 14079098880, "tok_s": 174939.3, "elapsed_s": 27005.1, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13460, "epoch": 0.904, "loss": 2.17571, "ppl": 8.808, "lr": 3.6592e-05, "lr_sched": "cosine", "grad_norm": 1.4635, "tokens": 14100049920, "tok_s": 174827.2, "elapsed_s": 27125.0, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13480, "epoch": 0.905, "loss": 2.14921, "ppl": 8.578, "lr": 3.6411e-05, "lr_sched": "cosine", "grad_norm": 1.4304, "tokens": 14121000960, "tok_s": 174318.9, "elapsed_s": 27244.9, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13500, "epoch": 0.906, "loss": 2.11365, "ppl": 8.278, "lr": 3.6233e-05, "lr_sched": "cosine", "grad_norm": 1.6806, "tokens": 14141952000, "tok_s": 174958.8, "elapsed_s": 27364.6, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13520, "epoch": 0.908, "loss": 2.24003, "ppl": 9.394, "lr": 3.6057e-05, "lr_sched": "cosine", "grad_norm": 1.5709, "tokens": 14162903040, "tok_s": 174778.6, "elapsed_s": 27484.4, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13540, "epoch": 0.909, "loss": 2.18752, "ppl": 8.913, "lr": 3.5884e-05, "lr_sched": "cosine", "grad_norm": 0.9345, "tokens": 14183854080, "tok_s": 174112.3, "elapsed_s": 27605.5, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13560, "epoch": 0.91, "loss": 2.16383, "ppl": 8.704, "lr": 3.5713e-05, "lr_sched": "cosine", "grad_norm": 0.9712, "tokens": 14204805120, "tok_s": 174529.4, "elapsed_s": 27725.9, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13580, "epoch": 0.912, "loss": 2.19781, "ppl": 9.005, "lr": 3.5545e-05, "lr_sched": "cosine", "grad_norm": 0.9147, "tokens": 14225756160, "tok_s": 174869.8, "elapsed_s": 27845.8, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13600, "epoch": 0.913, "loss": 2.19721, "ppl": 9.0, "lr": 3.5379e-05, "lr_sched": "cosine", "grad_norm": 0.8484, "tokens": 14246707200, "tok_s": 174332.4, "elapsed_s": 27965.7, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13600, "epoch": 0.913, "val_loss": 2.26266, "val_ppl": 9.609, "best_val_ppl": 9.468, "is_best": false}
{"step": 13620, "epoch": 0.914, "loss": 2.20698, "ppl": 9.088, "lr": 3.5216e-05, "lr_sched": "cosine", "grad_norm": 0.7151, "tokens": 14267658240, "tok_s": 173769.9, "elapsed_s": 28517.6, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13640, "epoch": 0.916, "loss": 2.16622, "ppl": 8.725, "lr": 3.5055e-05, "lr_sched": "cosine", "grad_norm": 0.8459, "tokens": 14288609280, "tok_s": 175331.8, "elapsed_s": 28637.6, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13660, "epoch": 0.917, "loss": 2.16298, "ppl": 8.697, "lr": 3.4896e-05, "lr_sched": "cosine", "grad_norm": 0.8043, "tokens": 14309560320, "tok_s": 174960.5, "elapsed_s": 28757.5, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13680, "epoch": 0.918, "loss": 2.15604, "ppl": 8.637, "lr": 3.474e-05, "lr_sched": "cosine", "grad_norm": 0.9858, "tokens": 14330511360, "tok_s": 172336.1, "elapsed_s": 28878.4, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13700, "epoch": 0.92, "loss": 2.07089, "ppl": 7.932, "lr": 3.4587e-05, "lr_sched": "cosine", "grad_norm": 0.9622, "tokens": 14351462400, "tok_s": 174751.4, "elapsed_s": 28998.7, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13720, "epoch": 0.921, "loss": 2.10553, "ppl": 8.211, "lr": 3.4436e-05, "lr_sched": "cosine", "grad_norm": 1.1568, "tokens": 14372413440, "tok_s": 172662.9, "elapsed_s": 29118.7, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13740, "epoch": 0.922, "loss": 2.2158, "ppl": 9.169, "lr": 3.4287e-05, "lr_sched": "cosine", "grad_norm": 0.8531, "tokens": 14393364480, "tok_s": 172788.4, "elapsed_s": 29240.0, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13760, "epoch": 0.924, "loss": 2.14107, "ppl": 8.509, "lr": 3.4141e-05, "lr_sched": "cosine", "grad_norm": 0.7686, "tokens": 14414315520, "tok_s": 174890.0, "elapsed_s": 29360.3, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13780, "epoch": 0.925, "loss": 2.15997, "ppl": 8.671, "lr": 3.3997e-05, "lr_sched": "cosine", "grad_norm": 0.9084, "tokens": 14435266560, "tok_s": 174370.4, "elapsed_s": 29480.3, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13800, "epoch": 0.926, "loss": 2.20769, "ppl": 9.095, "lr": 3.3856e-05, "lr_sched": "cosine", "grad_norm": 0.909, "tokens": 14456217600, "tok_s": 174939.1, "elapsed_s": 29600.2, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13820, "epoch": 0.928, "loss": 2.16714, "ppl": 8.733, "lr": 3.3718e-05, "lr_sched": "cosine", "grad_norm": 0.8645, "tokens": 14477168640, "tok_s": 174028.0, "elapsed_s": 29720.4, "vram_gb": 0.59, "ram_pct": 14.9, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13840, "epoch": 0.929, "loss": 2.05157, "ppl": 7.78, "lr": 3.3582e-05, "lr_sched": "cosine", "grad_norm": 0.9131, "tokens": 14498119680, "tok_s": 174829.3, "elapsed_s": 29840.6, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13860, "epoch": 0.93, "loss": 2.1585, "ppl": 8.658, "lr": 3.3448e-05, "lr_sched": "cosine", "grad_norm": 0.9685, "tokens": 14519070720, "tok_s": 174515.8, "elapsed_s": 29960.7, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13880, "epoch": 0.932, "loss": 2.20308, "ppl": 9.053, "lr": 3.3317e-05, "lr_sched": "cosine", "grad_norm": 0.8646, "tokens": 14540021760, "tok_s": 173617.2, "elapsed_s": 30080.9, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13900, "epoch": 0.933, "loss": 2.24449, "ppl": 9.436, "lr": 3.3189e-05, "lr_sched": "cosine", "grad_norm": 0.8958, "tokens": 14560972800, "tok_s": 174648.7, "elapsed_s": 30201.2, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13920, "epoch": 0.934, "loss": 2.26781, "ppl": 9.658, "lr": 3.3063e-05, "lr_sched": "cosine", "grad_norm": 0.8472, "tokens": 14581923840, "tok_s": 174015.8, "elapsed_s": 30321.3, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13940, "epoch": 0.936, "loss": 2.22956, "ppl": 9.296, "lr": 3.2939e-05, "lr_sched": "cosine", "grad_norm": 1.0371, "tokens": 14602874880, "tok_s": 173046.7, "elapsed_s": 30441.2, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13960, "epoch": 0.937, "loss": 2.235, "ppl": 9.346, "lr": 3.2818e-05, "lr_sched": "cosine", "grad_norm": 0.9667, "tokens": 14623825920, "tok_s": 174550.7, "elapsed_s": 30562.2, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 13980, "epoch": 0.938, "loss": 2.31074, "ppl": 10.082, "lr": 3.27e-05, "lr_sched": "cosine", "grad_norm": 1.0774, "tokens": 14644776960, "tok_s": 173442.5, "elapsed_s": 30682.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14000, "epoch": 0.94, "loss": 2.25429, "ppl": 9.529, "lr": 3.2584e-05, "lr_sched": "cosine", "grad_norm": 0.8264, "tokens": 14665728000, "tok_s": 174364.8, "elapsed_s": 30802.7, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14000, "epoch": 0.94, "val_loss": 2.24492, "val_ppl": 9.44, "best_val_ppl": 9.44, "is_best": true}
{"step": 14020, "epoch": 0.941, "loss": 2.22323, "ppl": 9.237, "lr": 3.247e-05, "lr_sched": "cosine", "grad_norm": 1.2341, "tokens": 14686679040, "tok_s": 173201.5, "elapsed_s": 31354.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14040, "epoch": 0.942, "loss": 2.29687, "ppl": 9.943, "lr": 3.2359e-05, "lr_sched": "cosine", "grad_norm": 0.9862, "tokens": 14707630080, "tok_s": 173384.1, "elapsed_s": 31474.6, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14060, "epoch": 0.944, "loss": 2.29389, "ppl": 9.913, "lr": 3.2251e-05, "lr_sched": "cosine", "grad_norm": 1.0619, "tokens": 14728581120, "tok_s": 174478.5, "elapsed_s": 31595.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14080, "epoch": 0.945, "loss": 2.25452, "ppl": 9.531, "lr": 3.2145e-05, "lr_sched": "cosine", "grad_norm": 0.8853, "tokens": 14749532160, "tok_s": 174636.2, "elapsed_s": 31715.5, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14100, "epoch": 0.946, "loss": 2.26522, "ppl": 9.633, "lr": 3.2042e-05, "lr_sched": "cosine", "grad_norm": 0.8977, "tokens": 14770483200, "tok_s": 174507.6, "elapsed_s": 31835.5, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14120, "epoch": 0.948, "loss": 2.2554, "ppl": 9.539, "lr": 3.1941e-05, "lr_sched": "cosine", "grad_norm": 0.9246, "tokens": 14791434240, "tok_s": 174620.7, "elapsed_s": 31955.6, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14140, "epoch": 0.949, "loss": 2.27109, "ppl": 9.69, "lr": 3.1843e-05, "lr_sched": "cosine", "grad_norm": 0.8236, "tokens": 14812385280, "tok_s": 174071.1, "elapsed_s": 32075.7, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14160, "epoch": 0.951, "loss": 2.21461, "ppl": 9.158, "lr": 3.1747e-05, "lr_sched": "cosine", "grad_norm": 1.0335, "tokens": 14833336320, "tok_s": 174818.6, "elapsed_s": 32195.9, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14180, "epoch": 0.952, "loss": 2.26497, "ppl": 9.631, "lr": 3.1654e-05, "lr_sched": "cosine", "grad_norm": 0.7495, "tokens": 14854287360, "tok_s": 173633.4, "elapsed_s": 32316.2, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14200, "epoch": 0.953, "loss": 2.22748, "ppl": 9.276, "lr": 3.1563e-05, "lr_sched": "cosine", "grad_norm": 1.0531, "tokens": 14875238400, "tok_s": 173416.1, "elapsed_s": 32436.6, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14220, "epoch": 0.955, "loss": 2.21711, "ppl": 9.181, "lr": 3.1475e-05, "lr_sched": "cosine", "grad_norm": 1.093, "tokens": 14896189440, "tok_s": 174411.2, "elapsed_s": 32556.9, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14240, "epoch": 0.956, "loss": 2.26404, "ppl": 9.622, "lr": 3.1389e-05, "lr_sched": "cosine", "grad_norm": 1.0866, "tokens": 14917140480, "tok_s": 174432.9, "elapsed_s": 32677.0, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14260, "epoch": 0.957, "loss": 2.24938, "ppl": 9.482, "lr": 3.1306e-05, "lr_sched": "cosine", "grad_norm": 0.8187, "tokens": 14938091520, "tok_s": 174929.9, "elapsed_s": 32797.1, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14280, "epoch": 0.959, "loss": 2.23802, "ppl": 9.375, "lr": 3.1226e-05, "lr_sched": "cosine", "grad_norm": 0.9001, "tokens": 14959042560, "tok_s": 174456.8, "elapsed_s": 32917.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14300, "epoch": 0.96, "loss": 2.20557, "ppl": 9.075, "lr": 3.1148e-05, "lr_sched": "cosine", "grad_norm": 0.8494, "tokens": 14979993600, "tok_s": 174670.9, "elapsed_s": 33037.3, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14320, "epoch": 0.961, "loss": 2.23598, "ppl": 9.356, "lr": 3.1072e-05, "lr_sched": "cosine", "grad_norm": 1.1121, "tokens": 15000944640, "tok_s": 174454.0, "elapsed_s": 33157.2, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14340, "epoch": 0.963, "loss": 2.26398, "ppl": 9.621, "lr": 3.1e-05, "lr_sched": "cosine", "grad_norm": 1.1046, "tokens": 15021895680, "tok_s": 174602.6, "elapsed_s": 33277.2, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14360, "epoch": 0.964, "loss": 2.25358, "ppl": 9.522, "lr": 3.0929e-05, "lr_sched": "cosine", "grad_norm": 1.1756, "tokens": 15042846720, "tok_s": 174376.9, "elapsed_s": 33397.2, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14380, "epoch": 0.965, "loss": 2.24429, "ppl": 9.434, "lr": 3.0862e-05, "lr_sched": "cosine", "grad_norm": 0.9133, "tokens": 15063797760, "tok_s": 174220.9, "elapsed_s": 33517.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14400, "epoch": 0.967, "loss": 2.23253, "ppl": 9.323, "lr": 3.0796e-05, "lr_sched": "cosine", "grad_norm": 0.9686, "tokens": 15084748800, "tok_s": 173673.2, "elapsed_s": 33637.7, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14400, "epoch": 0.967, "val_loss": 2.23857, "val_ppl": 9.38, "best_val_ppl": 9.38, "is_best": true}
{"step": 14420, "epoch": 0.968, "loss": 2.26988, "ppl": 9.678, "lr": 3.0734e-05, "lr_sched": "cosine", "grad_norm": 0.9488, "tokens": 15105699840, "tok_s": 172501.2, "elapsed_s": 34190.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14440, "epoch": 0.969, "loss": 2.24292, "ppl": 9.421, "lr": 3.0674e-05, "lr_sched": "cosine", "grad_norm": 0.7908, "tokens": 15126650880, "tok_s": 174680.4, "elapsed_s": 34311.4, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14460, "epoch": 0.971, "loss": 2.18435, "ppl": 8.885, "lr": 3.0616e-05, "lr_sched": "cosine", "grad_norm": 0.8013, "tokens": 15147601920, "tok_s": 174742.3, "elapsed_s": 34431.3, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14480, "epoch": 0.972, "loss": 2.27219, "ppl": 9.701, "lr": 3.0561e-05, "lr_sched": "cosine", "grad_norm": 0.8496, "tokens": 15168552960, "tok_s": 174349.9, "elapsed_s": 34551.4, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14500, "epoch": 0.973, "loss": 2.25377, "ppl": 9.524, "lr": 3.0509e-05, "lr_sched": "cosine", "grad_norm": 0.8429, "tokens": 15189504000, "tok_s": 174223.6, "elapsed_s": 34671.9, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14520, "epoch": 0.975, "loss": 2.17904, "ppl": 8.838, "lr": 3.0459e-05, "lr_sched": "cosine", "grad_norm": 1.1904, "tokens": 15210455040, "tok_s": 173967.0, "elapsed_s": 34792.5, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14540, "epoch": 0.976, "loss": 2.26488, "ppl": 9.63, "lr": 3.0412e-05, "lr_sched": "cosine", "grad_norm": 0.75, "tokens": 15231406080, "tok_s": 172685.4, "elapsed_s": 34913.0, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14560, "epoch": 0.977, "loss": 2.26077, "ppl": 9.59, "lr": 3.0367e-05, "lr_sched": "cosine", "grad_norm": 0.9646, "tokens": 15252357120, "tok_s": 173833.2, "elapsed_s": 35033.6, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14580, "epoch": 0.979, "loss": 2.18613, "ppl": 8.901, "lr": 3.0325e-05, "lr_sched": "cosine", "grad_norm": 1.1231, "tokens": 15273308160, "tok_s": 174439.9, "elapsed_s": 35153.9, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14600, "epoch": 0.98, "loss": 2.24045, "ppl": 9.398, "lr": 3.0285e-05, "lr_sched": "cosine", "grad_norm": 0.7996, "tokens": 15294259200, "tok_s": 173786.2, "elapsed_s": 35274.3, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14620, "epoch": 0.981, "loss": 2.23671, "ppl": 9.363, "lr": 3.0248e-05, "lr_sched": "cosine", "grad_norm": 0.9817, "tokens": 15315210240, "tok_s": 173854.9, "elapsed_s": 35394.6, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14640, "epoch": 0.983, "loss": 2.23457, "ppl": 9.343, "lr": 3.0214e-05, "lr_sched": "cosine", "grad_norm": 0.9, "tokens": 15336161280, "tok_s": 174437.6, "elapsed_s": 35514.9, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14660, "epoch": 0.984, "loss": 2.18966, "ppl": 8.932, "lr": 3.0182e-05, "lr_sched": "cosine", "grad_norm": 0.973, "tokens": 15357112320, "tok_s": 173277.4, "elapsed_s": 35635.6, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14680, "epoch": 0.985, "loss": 2.29466, "ppl": 9.921, "lr": 3.0153e-05, "lr_sched": "cosine", "grad_norm": 2.3577, "tokens": 15378063360, "tok_s": 173689.9, "elapsed_s": 35756.3, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14700, "epoch": 0.987, "loss": 2.27828, "ppl": 9.76, "lr": 3.0126e-05, "lr_sched": "cosine", "grad_norm": 1.473, "tokens": 15399014400, "tok_s": 173344.6, "elapsed_s": 35877.2, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14720, "epoch": 0.988, "loss": 2.23755, "ppl": 9.37, "lr": 3.0102e-05, "lr_sched": "cosine", "grad_norm": 0.9834, "tokens": 15419965440, "tok_s": 174269.0, "elapsed_s": 35997.9, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14740, "epoch": 0.989, "loss": 2.12897, "ppl": 8.406, "lr": 3.008e-05, "lr_sched": "cosine", "grad_norm": 1.4754, "tokens": 15440916480, "tok_s": 173869.3, "elapsed_s": 36118.3, "vram_gb": 0.59, "ram_pct": 15.0, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14760, "epoch": 0.991, "loss": 2.18611, "ppl": 8.901, "lr": 3.0061e-05, "lr_sched": "cosine", "grad_norm": 1.3515, "tokens": 15461867520, "tok_s": 173718.8, "elapsed_s": 36238.7, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14780, "epoch": 0.992, "loss": 2.16559, "ppl": 8.72, "lr": 3.0045e-05, "lr_sched": "cosine", "grad_norm": 1.2611, "tokens": 15482818560, "tok_s": 174903.3, "elapsed_s": 36358.8, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14800, "epoch": 0.993, "loss": 2.1459, "ppl": 8.55, "lr": 3.0031e-05, "lr_sched": "cosine", "grad_norm": 0.9891, "tokens": 15503769600, "tok_s": 175021.1, "elapsed_s": 36478.7, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14800, "epoch": 0.993, "val_loss": 2.30844, "val_ppl": 10.059, "best_val_ppl": 9.38, "is_best": false}
{"step": 14820, "epoch": 0.995, "loss": 2.14151, "ppl": 8.512, "lr": 3.002e-05, "lr_sched": "cosine", "grad_norm": 1.0925, "tokens": 15524720640, "tok_s": 173867.4, "elapsed_s": 37032.0, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14840, "epoch": 0.996, "loss": 2.06561, "ppl": 7.89, "lr": 3.0011e-05, "lr_sched": "cosine", "grad_norm": 1.7626, "tokens": 15545671680, "tok_s": 173214.4, "elapsed_s": 37152.2, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14860, "epoch": 0.998, "loss": 2.22387, "ppl": 9.243, "lr": 3.0005e-05, "lr_sched": "cosine", "grad_norm": 1.1266, "tokens": 15566622720, "tok_s": 174985.1, "elapsed_s": 37272.0, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}
{"step": 14880, "epoch": 0.999, "loss": 2.17615, "ppl": 8.812, "lr": 3.0001e-05, "lr_sched": "cosine", "grad_norm": 1.0174, "tokens": 15587573760, "tok_s": 174865.8, "elapsed_s": 37392.1, "vram_gb": 0.59, "ram_pct": 15.1, "disk_free_gb": 20.8, "attn_backend": "chunked", "amp_dtype": "torch.float16"}