Hugging Face's logo

Reverent
/
llama3-8b-code-extended

Safetensors
llama
Model card Files
xet
 Community
Reverent commited on
Commit
f87d4d3
·
verified ·
1 Parent(s): 0cc1630

Training in progress, step 5000

Browse files
Files changed (3) hide show
  1. model.safetensors +1 -1
  2. training_args.bin +1 -1
  3. training_log.txt +503 -402
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:635e738f919678189191de44fd4b70faf93107154227cb6e9c11c84668e539a1
3
  size 16060556616
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c24eefb159064476b2be5d5a5a23bd99c7941b07ca3fc1e6e96cfee0ea3bfaa2
3
  size 16060556616
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:57d5c33f56445706ad779b961edb4396188b7b460b6b372abc6a8f229db21776
3
  size 5265
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:566b98d618c43899f8b01538667552d7c8a4f9f3959066cd81b19bd535ee9706
3
  size 5265
training_log.txt CHANGED
@@ -1,406 +1,507 @@
1
 
2
  ==================================================
3
- Training started at: 2026-04-04 09:18:44
4
  ==================================================
5
 
6
- [2026-04-04 09:24:56] Step 2010: loss: 2.8399, grad_norm: 1.7500, learning_rate: 0.0000, epoch: 0.0100
7
- [2026-04-04 09:26:20] Step 2020: loss: 2.5400, grad_norm: 3.1562, learning_rate: 0.0000, epoch: 0.0101
8
- [2026-04-04 09:27:44] Step 2030: loss: 2.2588, grad_norm: 0.6328, learning_rate: 0.0000, epoch: 0.0101
9
- [2026-04-04 09:29:08] Step 2040: loss: 2.5586, grad_norm: 2.8906, learning_rate: 0.0000, epoch: 0.0102
10
- [2026-04-04 09:30:32] Step 2050: loss: 1.3992, grad_norm: 0.3848, learning_rate: 0.0000, epoch: 0.0103
11
- [2026-04-04 09:31:57] Step 2060: loss: 0.9564, grad_norm: 0.6211, learning_rate: 0.0000, epoch: 0.0103
12
- [2026-04-04 09:33:21] Step 2070: loss: 1.0082, grad_norm: 0.3125, learning_rate: 0.0000, epoch: 0.0103
13
- [2026-04-04 09:34:45] Step 2080: loss: 1.0009, grad_norm: 0.3047, learning_rate: 0.0000, epoch: 0.0104
14
- [2026-04-04 09:36:10] Step 2090: loss: 1.0767, grad_norm: 4.9375, learning_rate: 0.0000, epoch: 0.0104
15
- [2026-04-04 09:37:34] Step 2100: loss: 1.2326, grad_norm: 7.1562, learning_rate: 0.0000, epoch: 0.0105
16
- [2026-04-04 09:38:58] Step 2110: loss: 1.0480, grad_norm: 0.3496, learning_rate: 0.0000, epoch: 0.0106
17
- [2026-04-04 09:40:22] Step 2120: loss: 0.9334, grad_norm: 1.3828, learning_rate: 0.0000, epoch: 0.0106
18
- [2026-04-04 09:41:46] Step 2130: loss: 0.9686, grad_norm: 0.3398, learning_rate: 0.0000, epoch: 0.0106
19
- [2026-04-04 09:43:10] Step 2140: loss: 0.9875, grad_norm: 0.2695, learning_rate: 0.0000, epoch: 0.0107
20
- [2026-04-04 09:44:34] Step 2150: loss: 1.0286, grad_norm: 0.4629, learning_rate: 0.0000, epoch: 0.0107
21
- [2026-04-04 09:45:59] Step 2160: loss: 0.9684, grad_norm: 0.4922, learning_rate: 0.0000, epoch: 0.0108
22
- [2026-04-04 09:47:23] Step 2170: loss: 1.0229, grad_norm: 0.3691, learning_rate: 0.0000, epoch: 0.0109
23
- [2026-04-04 09:48:48] Step 2180: loss: 0.9697, grad_norm: 0.2598, learning_rate: 0.0000, epoch: 0.0109
24
- [2026-04-04 09:50:12] Step 2190: loss: 0.9916, grad_norm: 0.2852, learning_rate: 0.0000, epoch: 0.0109
25
- [2026-04-04 09:51:36] Step 2200: loss: 0.7603, grad_norm: 0.3242, learning_rate: 0.0000, epoch: 0.0110
26
- [2026-04-04 09:53:00] Step 2210: loss: 0.8609, grad_norm: 0.2617, learning_rate: 0.0000, epoch: 0.0111
27
- [2026-04-04 09:54:24] Step 2220: loss: 0.9794, grad_norm: 0.2363, learning_rate: 0.0000, epoch: 0.0111
28
- [2026-04-04 09:55:48] Step 2230: loss: 0.8372, grad_norm: 0.2451, learning_rate: 0.0000, epoch: 0.0112
29
- [2026-04-04 09:57:13] Step 2240: loss: 1.0379, grad_norm: 0.2090, learning_rate: 0.0000, epoch: 0.0112
30
- [2026-04-04 09:58:37] Step 2250: loss: 0.9140, grad_norm: 0.5703, learning_rate: 0.0000, epoch: 0.0112
31
- [2026-04-04 10:00:02] Step 2260: loss: 0.9224, grad_norm: 0.2041, learning_rate: 0.0000, epoch: 0.0113
32
- [2026-04-04 10:01:26] Step 2270: loss: 0.9180, grad_norm: 0.2061, learning_rate: 0.0000, epoch: 0.0114
33
- [2026-04-04 10:02:50] Step 2280: loss: 0.9332, grad_norm: 0.2324, learning_rate: 0.0000, epoch: 0.0114
34
- [2026-04-04 10:04:15] Step 2290: loss: 1.0547, grad_norm: 0.2490, learning_rate: 0.0000, epoch: 0.0115
35
- [2026-04-04 10:05:38] Step 2300: loss: 0.9023, grad_norm: 0.2402, learning_rate: 0.0000, epoch: 0.0115
36
- [2026-04-04 10:07:03] Step 2310: loss: 0.9282, grad_norm: 0.2471, learning_rate: 0.0000, epoch: 0.0115
37
- [2026-04-04 10:08:27] Step 2320: loss: 1.0646, grad_norm: 0.3613, learning_rate: 0.0000, epoch: 0.0116
38
- [2026-04-04 10:09:51] Step 2330: loss: 1.0843, grad_norm: 0.2295, learning_rate: 0.0000, epoch: 0.0117
39
- [2026-04-04 10:11:15] Step 2340: loss: 0.8962, grad_norm: 0.4238, learning_rate: 0.0000, epoch: 0.0117
40
- [2026-04-04 10:12:39] Step 2350: loss: 1.1290, grad_norm: 0.2246, learning_rate: 0.0000, epoch: 0.0118
41
- [2026-04-04 10:14:03] Step 2360: loss: 0.8712, grad_norm: 0.3027, learning_rate: 0.0000, epoch: 0.0118
42
- [2026-04-04 10:15:28] Step 2370: loss: 0.8248, grad_norm: 0.1875, learning_rate: 0.0000, epoch: 0.0118
43
- [2026-04-04 10:16:52] Step 2380: loss: 0.8929, grad_norm: 0.2314, learning_rate: 0.0000, epoch: 0.0119
44
- [2026-04-04 10:18:16] Step 2390: loss: 0.9730, grad_norm: 0.2129, learning_rate: 0.0000, epoch: 0.0120
45
- [2026-04-04 10:19:40] Step 2400: loss: 0.8794, grad_norm: 0.3340, learning_rate: 0.0000, epoch: 0.0120
46
- [2026-04-04 10:21:04] Step 2410: loss: 0.8079, grad_norm: 0.2637, learning_rate: 0.0000, epoch: 0.0120
47
- [2026-04-04 10:22:29] Step 2420: loss: 0.9668, grad_norm: 0.2490, learning_rate: 0.0000, epoch: 0.0121
48
- [2026-04-04 10:23:53] Step 2430: loss: 0.8787, grad_norm: 0.2969, learning_rate: 0.0000, epoch: 0.0121
49
- [2026-04-04 10:25:17] Step 2440: loss: 0.8943, grad_norm: 0.2852, learning_rate: 0.0000, epoch: 0.0122
50
- [2026-04-04 10:26:41] Step 2450: loss: 0.9024, grad_norm: 0.2637, learning_rate: 0.0000, epoch: 0.0123
51
- [2026-04-04 10:28:05] Step 2460: loss: 0.7944, grad_norm: 0.3672, learning_rate: 0.0000, epoch: 0.0123
52
- [2026-04-04 10:29:30] Step 2470: loss: 0.9620, grad_norm: 0.2109, learning_rate: 0.0000, epoch: 0.0123
53
- [2026-04-04 10:30:54] Step 2480: loss: 0.8795, grad_norm: 0.2637, learning_rate: 0.0000, epoch: 0.0124
54
- [2026-04-04 10:32:19] Step 2490: loss: 0.8035, grad_norm: 0.2295, learning_rate: 0.0000, epoch: 0.0124
55
- [2026-04-04 10:33:43] Step 2500: loss: 0.7863, grad_norm: 0.3164, learning_rate: 0.0000, epoch: 0.0125
56
- [2026-04-04 10:35:07] Step 2510: loss: 1.0078, grad_norm: 0.5156, learning_rate: 0.0000, epoch: 0.0126
57
- [2026-04-04 10:36:32] Step 2520: loss: 0.8982, grad_norm: 0.2559, learning_rate: 0.0000, epoch: 0.0126
58
- [2026-04-04 10:37:56] Step 2530: loss: 0.9230, grad_norm: 0.2070, learning_rate: 0.0000, epoch: 0.0126
59
- [2026-04-04 10:39:19] Step 2540: loss: 0.4966, grad_norm: 0.2969, learning_rate: 0.0000, epoch: 0.0127
60
- [2026-04-04 10:40:43] Step 2550: loss: 0.9865, grad_norm: 0.2305, learning_rate: 0.0000, epoch: 0.0127
61
- [2026-04-04 10:42:07] Step 2560: loss: 0.8903, grad_norm: 0.2402, learning_rate: 0.0000, epoch: 0.0128
62
- [2026-04-04 10:43:31] Step 2570: loss: 0.9468, grad_norm: 0.4883, learning_rate: 0.0000, epoch: 0.0129
63
- [2026-04-04 10:44:56] Step 2580: loss: 0.9083, grad_norm: 0.2490, learning_rate: 0.0000, epoch: 0.0129
64
- [2026-04-04 10:46:20] Step 2590: loss: 0.9190, grad_norm: 0.2041, learning_rate: 0.0000, epoch: 0.0129
65
- [2026-04-04 10:47:44] Step 2600: loss: 0.9126, grad_norm: 0.2373, learning_rate: 0.0000, epoch: 0.0130
66
- [2026-04-04 10:49:09] Step 2610: loss: 0.8419, grad_norm: 0.3125, learning_rate: 0.0000, epoch: 0.0131
67
- [2026-04-04 10:50:33] Step 2620: loss: 0.7891, grad_norm: 0.1816, learning_rate: 0.0000, epoch: 0.0131
68
- [2026-04-04 10:51:57] Step 2630: loss: 0.8950, grad_norm: 0.2314, learning_rate: 0.0000, epoch: 0.0132
69
- [2026-04-04 10:53:21] Step 2640: loss: 0.8428, grad_norm: 0.2539, learning_rate: 0.0000, epoch: 0.0132
70
- [2026-04-04 10:54:45] Step 2650: loss: 0.9133, grad_norm: 0.2734, learning_rate: 0.0000, epoch: 0.0132
71
- [2026-04-04 10:56:09] Step 2660: loss: 1.2072, grad_norm: 0.2793, learning_rate: 0.0000, epoch: 0.0133
72
- [2026-04-04 10:57:34] Step 2670: loss: 0.9179, grad_norm: 1.3750, learning_rate: 0.0000, epoch: 0.0134
73
- [2026-04-04 10:58:58] Step 2680: loss: 0.8475, grad_norm: 0.2266, learning_rate: 0.0000, epoch: 0.0134
74
- [2026-04-04 11:00:22] Step 2690: loss: 0.8865, grad_norm: 0.3066, learning_rate: 0.0000, epoch: 0.0135
75
- [2026-04-04 11:01:47] Step 2700: loss: 1.1335, grad_norm: 0.2578, learning_rate: 0.0000, epoch: 0.0135
76
- [2026-04-04 11:03:11] Step 2710: loss: 0.8892, grad_norm: 0.3223, learning_rate: 0.0000, epoch: 0.0135
77
- [2026-04-04 11:04:35] Step 2720: loss: 0.8915, grad_norm: 0.3145, learning_rate: 0.0000, epoch: 0.0136
78
- [2026-04-04 11:06:00] Step 2730: loss: 0.8971, grad_norm: 0.2383, learning_rate: 0.0000, epoch: 0.0137
79
- [2026-04-04 11:07:24] Step 2740: loss: 0.9294, grad_norm: 0.2695, learning_rate: 0.0000, epoch: 0.0137
80
- [2026-04-04 11:08:48] Step 2750: loss: 0.8750, grad_norm: 0.1738, learning_rate: 0.0000, epoch: 0.0138
81
- [2026-04-04 11:10:12] Step 2760: loss: 0.8538, grad_norm: 0.9219, learning_rate: 0.0000, epoch: 0.0138
82
- [2026-04-04 11:11:36] Step 2770: loss: 0.8476, grad_norm: 0.2754, learning_rate: 0.0000, epoch: 0.0138
83
- [2026-04-04 11:13:00] Step 2780: loss: 0.7105, grad_norm: 0.2656, learning_rate: 0.0000, epoch: 0.0139
84
- [2026-04-04 11:14:24] Step 2790: loss: 0.8698, grad_norm: 0.5781, learning_rate: 0.0000, epoch: 0.0140
85
- [2026-04-04 11:15:48] Step 2800: loss: 0.9280, grad_norm: 0.2305, learning_rate: 0.0000, epoch: 0.0140
86
- [2026-04-04 11:17:12] Step 2810: loss: 0.9404, grad_norm: 0.2236, learning_rate: 0.0000, epoch: 0.0140
87
- [2026-04-04 11:18:36] Step 2820: loss: 0.8950, grad_norm: 0.2188, learning_rate: 0.0000, epoch: 0.0141
88
- [2026-04-04 11:20:00] Step 2830: loss: 0.9473, grad_norm: 1.7734, learning_rate: 0.0000, epoch: 0.0141
89
- [2026-04-04 11:21:24] Step 2840: loss: 0.8666, grad_norm: 0.1504, learning_rate: 0.0000, epoch: 0.0142
90
- [2026-04-04 11:22:48] Step 2850: loss: 0.9285, grad_norm: 0.2275, learning_rate: 0.0000, epoch: 0.0143
91
- [2026-04-04 11:24:12] Step 2860: loss: 0.8818, grad_norm: 0.2598, learning_rate: 0.0000, epoch: 0.0143
92
- [2026-04-04 11:25:36] Step 2870: loss: 0.8695, grad_norm: 0.3086, learning_rate: 0.0000, epoch: 0.0143
93
- [2026-04-04 11:27:00] Step 2880: loss: 0.8156, grad_norm: 0.2432, learning_rate: 0.0000, epoch: 0.0144
94
- [2026-04-04 11:28:24] Step 2890: loss: 1.0572, grad_norm: 0.2012, learning_rate: 0.0000, epoch: 0.0144
95
- [2026-04-04 11:29:48] Step 2900: loss: 0.9074, grad_norm: 0.3926, learning_rate: 0.0000, epoch: 0.0145
96
- [2026-04-04 11:31:12] Step 2910: loss: 0.9228, grad_norm: 0.3008, learning_rate: 0.0000, epoch: 0.0146
97
- [2026-04-04 11:32:36] Step 2920: loss: 0.7711, grad_norm: 0.2148, learning_rate: 0.0000, epoch: 0.0146
98
- [2026-04-04 11:33:59] Step 2930: loss: 0.8896, grad_norm: 0.1621, learning_rate: 0.0000, epoch: 0.0146
99
- [2026-04-04 11:35:24] Step 2940: loss: 0.8782, grad_norm: 0.2031, learning_rate: 0.0000, epoch: 0.0147
100
- [2026-04-04 11:36:47] Step 2950: loss: 0.8114, grad_norm: 0.3398, learning_rate: 0.0000, epoch: 0.0147
101
- [2026-04-04 11:38:12] Step 2960: loss: 0.8362, grad_norm: 0.1992, learning_rate: 0.0000, epoch: 0.0148
102
- [2026-04-04 11:39:36] Step 2970: loss: 0.9280, grad_norm: 0.2402, learning_rate: 0.0000, epoch: 0.0149
103
- [2026-04-04 11:41:00] Step 2980: loss: 0.8748, grad_norm: 0.1982, learning_rate: 0.0000, epoch: 0.0149
104
- [2026-04-04 11:42:24] Step 2990: loss: 0.8827, grad_norm: 0.2451, learning_rate: 0.0000, epoch: 0.0149
105
- [2026-04-04 11:43:48] Step 3000: loss: 0.8786, grad_norm: 0.6719, learning_rate: 0.0000, epoch: 0.0150
106
- [2026-04-04 11:45:12] Step 3010: loss: 0.9199, grad_norm: 10.9375, learning_rate: 0.0000, epoch: 0.0150
107
- [2026-04-04 11:46:36] Step 3020: loss: 0.9592, grad_norm: 0.3965, learning_rate: 0.0000, epoch: 0.0151
108
- [2026-04-04 11:48:00] Step 3030: loss: 0.9399, grad_norm: 0.2031, learning_rate: 0.0000, epoch: 0.0152
109
- [2026-04-04 11:49:24] Step 3040: loss: 0.9260, grad_norm: 0.2617, learning_rate: 0.0000, epoch: 0.0152
110
- [2026-04-04 11:50:48] Step 3050: loss: 0.8369, grad_norm: 0.2695, learning_rate: 0.0000, epoch: 0.0152
111
- [2026-04-04 11:52:13] Step 3060: loss: 0.8429, grad_norm: 0.2285, learning_rate: 0.0000, epoch: 0.0153
112
- [2026-04-04 11:53:37] Step 3070: loss: 0.9259, grad_norm: 0.2637, learning_rate: 0.0000, epoch: 0.0154
113
- [2026-04-04 11:55:01] Step 3080: loss: 0.9881, grad_norm: 0.2412, learning_rate: 0.0000, epoch: 0.0154
114
- [2026-04-04 11:56:25] Step 3090: loss: 0.9997, grad_norm: 0.2207, learning_rate: 0.0000, epoch: 0.0155
115
- [2026-04-04 11:57:49] Step 3100: loss: 0.8810, grad_norm: 0.4238, learning_rate: 0.0000, epoch: 0.0155
116
- [2026-04-04 11:59:14] Step 3110: loss: 0.8769, grad_norm: 0.2500, learning_rate: 0.0000, epoch: 0.0155
117
- [2026-04-04 12:00:39] Step 3120: loss: 0.8588, grad_norm: 0.2598, learning_rate: 0.0000, epoch: 0.0156
118
- [2026-04-04 12:02:03] Step 3130: loss: 0.8142, grad_norm: 0.1787, learning_rate: 0.0000, epoch: 0.0157
119
- [2026-04-04 12:03:28] Step 3140: loss: 0.8229, grad_norm: 0.1855, learning_rate: 0.0000, epoch: 0.0157
120
- [2026-04-04 12:04:52] Step 3150: loss: 0.8981, grad_norm: 0.1797, learning_rate: 0.0000, epoch: 0.0158
121
- [2026-04-04 12:06:17] Step 3160: loss: 0.9945, grad_norm: 0.2109, learning_rate: 0.0000, epoch: 0.0158
122
- [2026-04-04 12:07:42] Step 3170: loss: 0.8905, grad_norm: 0.1758, learning_rate: 0.0000, epoch: 0.0158
123
- [2026-04-04 12:09:07] Step 3180: loss: 0.8941, grad_norm: 0.1865, learning_rate: 0.0000, epoch: 0.0159
124
- [2026-04-04 12:10:32] Step 3190: loss: 0.8214, grad_norm: 0.1826, learning_rate: 0.0000, epoch: 0.0159
125
- [2026-04-04 12:11:56] Step 3200: loss: 0.8866, grad_norm: 0.1865, learning_rate: 0.0000, epoch: 0.0160
126
- [2026-04-04 12:13:20] Step 3210: loss: 0.7030, grad_norm: 0.2178, learning_rate: 0.0000, epoch: 0.0160
127
- [2026-04-04 12:14:45] Step 3220: loss: 0.8260, grad_norm: 0.1631, learning_rate: 0.0000, epoch: 0.0161
128
- [2026-04-04 12:16:09] Step 3230: loss: 0.8487, grad_norm: 0.1768, learning_rate: 0.0000, epoch: 0.0162
129
- [2026-04-04 12:17:34] Step 3240: loss: 0.8839, grad_norm: 1.5859, learning_rate: 0.0000, epoch: 0.0162
130
- [2026-04-04 12:18:59] Step 3250: loss: 0.8126, grad_norm: 0.1729, learning_rate: 0.0000, epoch: 0.0163
131
- [2026-04-04 12:20:22] Step 3260: loss: 0.8689, grad_norm: 3.5781, learning_rate: 0.0000, epoch: 0.0163
132
- [2026-04-04 12:21:46] Step 3270: loss: 0.8413, grad_norm: 0.7891, learning_rate: 0.0000, epoch: 0.0163
133
- [2026-04-04 12:23:11] Step 3280: loss: 0.9708, grad_norm: 0.4590, learning_rate: 0.0000, epoch: 0.0164
134
- [2026-04-04 12:24:35] Step 3290: loss: 0.8723, grad_norm: 0.2754, learning_rate: 0.0000, epoch: 0.0164
135
- [2026-04-04 12:26:00] Step 3300: loss: 0.9940, grad_norm: 0.2246, learning_rate: 0.0000, epoch: 0.0165
136
- [2026-04-04 12:27:24] Step 3310: loss: 0.8340, grad_norm: 1.4609, learning_rate: 0.0000, epoch: 0.0165
137
- [2026-04-04 12:28:49] Step 3320: loss: 0.8388, grad_norm: 0.1953, learning_rate: 0.0000, epoch: 0.0166
138
- [2026-04-04 12:30:14] Step 3330: loss: 0.8837, grad_norm: 0.2148, learning_rate: 0.0000, epoch: 0.0167
139
- [2026-04-04 12:31:38] Step 3340: loss: 0.8853, grad_norm: 0.2354, learning_rate: 0.0000, epoch: 0.0167
140
- [2026-04-04 12:33:02] Step 3350: loss: 0.9380, grad_norm: 0.2168, learning_rate: 0.0000, epoch: 0.0168
141
- [2026-04-04 12:34:27] Step 3360: loss: 0.8515, grad_norm: 0.2021, learning_rate: 0.0000, epoch: 0.0168
142
- [2026-04-04 12:35:51] Step 3370: loss: 0.9047, grad_norm: 0.1885, learning_rate: 0.0000, epoch: 0.0169
143
- [2026-04-04 12:37:16] Step 3380: loss: 0.7929, grad_norm: 1.0000, learning_rate: 0.0000, epoch: 0.0169
144
- [2026-04-04 12:38:41] Step 3390: loss: 0.8338, grad_norm: 0.2012, learning_rate: 0.0000, epoch: 0.0169
145
- [2026-04-04 12:40:05] Step 3400: loss: 0.8262, grad_norm: 0.2559, learning_rate: 0.0000, epoch: 0.0170
146
- [2026-04-04 12:41:30] Step 3410: loss: 0.7306, grad_norm: 0.2295, learning_rate: 0.0000, epoch: 0.0170
147
- [2026-04-04 12:42:55] Step 3420: loss: 0.9324, grad_norm: 0.2520, learning_rate: 0.0000, epoch: 0.0171
148
- [2026-04-04 12:44:19] Step 3430: loss: 0.7892, grad_norm: 0.1533, learning_rate: 0.0000, epoch: 0.0171
149
- [2026-04-04 12:45:43] Step 3440: loss: 0.9207, grad_norm: 0.2139, learning_rate: 0.0000, epoch: 0.0172
150
- [2026-04-04 12:47:07] Step 3450: loss: 0.8582, grad_norm: 0.2539, learning_rate: 0.0000, epoch: 0.0173
151
- [2026-04-04 12:48:32] Step 3460: loss: 0.8868, grad_norm: 0.3867, learning_rate: 0.0000, epoch: 0.0173
152
- [2026-04-04 12:49:57] Step 3470: loss: 0.8362, grad_norm: 0.2148, learning_rate: 0.0000, epoch: 0.0174
153
- [2026-04-04 12:51:21] Step 3480: loss: 0.9078, grad_norm: 0.1855, learning_rate: 0.0000, epoch: 0.0174
154
- [2026-04-04 12:52:45] Step 3490: loss: 0.8814, grad_norm: 0.1660, learning_rate: 0.0000, epoch: 0.0175
155
- [2026-04-04 12:54:10] Step 3500: loss: 0.8289, grad_norm: 0.1992, learning_rate: 0.0000, epoch: 0.0175
156
- [2026-04-04 12:55:35] Step 3510: loss: 0.7498, grad_norm: 0.3477, learning_rate: 0.0000, epoch: 0.0175
157
- [2026-04-04 12:56:59] Step 3520: loss: 0.8706, grad_norm: 0.1504, learning_rate: 0.0000, epoch: 0.0176
158
- [2026-04-04 12:58:24] Step 3530: loss: 0.8755, grad_norm: 0.2480, learning_rate: 0.0000, epoch: 0.0176
159
- [2026-04-04 12:59:48] Step 3540: loss: 0.8941, grad_norm: 0.2041, learning_rate: 0.0000, epoch: 0.0177
160
- [2026-04-04 13:01:12] Step 3550: loss: 0.9034, grad_norm: 0.1914, learning_rate: 0.0000, epoch: 0.0177
161
- [2026-04-04 13:02:37] Step 3560: loss: 0.8194, grad_norm: 0.1680, learning_rate: 0.0000, epoch: 0.0178
162
- [2026-04-04 13:04:01] Step 3570: loss: 0.8267, grad_norm: 0.1807, learning_rate: 0.0000, epoch: 0.0179
163
- [2026-04-04 13:05:26] Step 3580: loss: 0.8188, grad_norm: 0.1445, learning_rate: 0.0000, epoch: 0.0179
164
- [2026-04-04 13:06:51] Step 3590: loss: 0.8666, grad_norm: 0.1797, learning_rate: 0.0000, epoch: 0.0180
165
- [2026-04-04 13:08:15] Step 3600: loss: 0.8775, grad_norm: 0.3203, learning_rate: 0.0000, epoch: 0.0180
166
- [2026-04-04 13:09:40] Step 3610: loss: 0.7504, grad_norm: 0.2734, learning_rate: 0.0000, epoch: 0.0181
167
- [2026-04-04 13:11:04] Step 3620: loss: 0.7219, grad_norm: 0.2393, learning_rate: 0.0000, epoch: 0.0181
168
- [2026-04-04 13:12:28] Step 3630: loss: 0.8402, grad_norm: 0.3105, learning_rate: 0.0000, epoch: 0.0181
169
- [2026-04-04 13:13:52] Step 3640: loss: 0.7577, grad_norm: 0.2559, learning_rate: 0.0000, epoch: 0.0182
170
- [2026-04-04 13:15:16] Step 3650: loss: 0.8412, grad_norm: 0.3867, learning_rate: 0.0000, epoch: 0.0182
171
- [2026-04-04 13:16:41] Step 3660: loss: 0.7960, grad_norm: 0.1533, learning_rate: 0.0000, epoch: 0.0183
172
- [2026-04-04 13:18:05] Step 3670: loss: 0.9164, grad_norm: 0.1787, learning_rate: 0.0000, epoch: 0.0184
173
- [2026-04-04 13:19:30] Step 3680: loss: 0.9855, grad_norm: 0.1270, learning_rate: 0.0000, epoch: 0.0184
174
- [2026-04-04 13:20:54] Step 3690: loss: 0.8332, grad_norm: 0.1455, learning_rate: 0.0000, epoch: 0.0185
175
- [2026-04-04 13:22:19] Step 3700: loss: 0.8512, grad_norm: 0.1924, learning_rate: 0.0000, epoch: 0.0185
176
- [2026-04-04 13:23:43] Step 3710: loss: 0.8994, grad_norm: 0.4375, learning_rate: 0.0000, epoch: 0.0186
177
- [2026-04-04 13:25:08] Step 3720: loss: 0.9205, grad_norm: 0.2314, learning_rate: 0.0000, epoch: 0.0186
178
- [2026-04-04 13:26:32] Step 3730: loss: 0.8814, grad_norm: 0.2227, learning_rate: 0.0000, epoch: 0.0186
179
- [2026-04-04 13:27:56] Step 3740: loss: 0.8907, grad_norm: 0.2227, learning_rate: 0.0000, epoch: 0.0187
180
- [2026-04-04 13:29:20] Step 3750: loss: 0.8737, grad_norm: 0.1475, learning_rate: 0.0000, epoch: 0.0187
181
- [2026-04-04 13:30:45] Step 3760: loss: 0.8853, grad_norm: 0.1943, learning_rate: 0.0000, epoch: 0.0188
182
- [2026-04-04 13:32:09] Step 3770: loss: 0.8969, grad_norm: 0.1943, learning_rate: 0.0000, epoch: 0.0188
183
- [2026-04-04 13:33:34] Step 3780: loss: 0.8080, grad_norm: 0.1992, learning_rate: 0.0000, epoch: 0.0189
184
- [2026-04-04 13:34:57] Step 3790: loss: 0.7392, grad_norm: 0.1592, learning_rate: 0.0000, epoch: 0.0190
185
- [2026-04-04 13:36:21] Step 3800: loss: 0.6776, grad_norm: 0.1875, learning_rate: 0.0000, epoch: 0.0190
186
- [2026-04-04 13:37:46] Step 3810: loss: 0.8414, grad_norm: 0.1836, learning_rate: 0.0000, epoch: 0.0191
187
- [2026-04-04 13:39:10] Step 3820: loss: 0.8751, grad_norm: 0.1699, learning_rate: 0.0000, epoch: 0.0191
188
- [2026-04-04 13:40:34] Step 3830: loss: 0.8723, grad_norm: 0.1738, learning_rate: 0.0000, epoch: 0.0192
189
- [2026-04-04 13:41:59] Step 3840: loss: 0.7960, grad_norm: 0.1758, learning_rate: 0.0000, epoch: 0.0192
190
- [2026-04-04 13:43:23] Step 3850: loss: 0.9287, grad_norm: 0.1865, learning_rate: 0.0000, epoch: 0.0192
191
- [2026-04-04 13:44:48] Step 3860: loss: 0.8020, grad_norm: 0.1934, learning_rate: 0.0000, epoch: 0.0193
192
- [2026-04-04 13:46:12] Step 3870: loss: 0.9291, grad_norm: 0.1758, learning_rate: 0.0000, epoch: 0.0193
193
- [2026-04-04 13:47:36] Step 3880: loss: 0.8263, grad_norm: 0.1816, learning_rate: 0.0000, epoch: 0.0194
194
- [2026-04-04 13:49:01] Step 3890: loss: 0.8684, grad_norm: 0.1641, learning_rate: 0.0000, epoch: 0.0194
195
- [2026-04-04 13:50:26] Step 3900: loss: 0.8842, grad_norm: 0.2363, learning_rate: 0.0000, epoch: 0.0195
196
- [2026-04-04 13:51:50] Step 3910: loss: 0.9238, grad_norm: 0.1885, learning_rate: 0.0000, epoch: 0.0196
197
- [2026-04-04 13:53:15] Step 3920: loss: 0.8393, grad_norm: 0.1748, learning_rate: 0.0000, epoch: 0.0196
198
- [2026-04-04 13:54:40] Step 3930: loss: 0.8357, grad_norm: 0.1436, learning_rate: 0.0000, epoch: 0.0197
199
- [2026-04-04 13:56:04] Step 3940: loss: 0.8333, grad_norm: 0.2021, learning_rate: 0.0000, epoch: 0.0197
200
- [2026-04-04 13:57:28] Step 3950: loss: 0.9847, grad_norm: 0.2031, learning_rate: 0.0000, epoch: 0.0198
201
- [2026-04-04 13:58:52] Step 3960: loss: 0.8540, grad_norm: 0.1670, learning_rate: 0.0000, epoch: 0.0198
202
- [2026-04-04 14:00:17] Step 3970: loss: 0.7641, grad_norm: 0.1436, learning_rate: 0.0000, epoch: 0.0198
203
- [2026-04-04 14:01:41] Step 3980: loss: 0.7934, grad_norm: 0.1621, learning_rate: 0.0000, epoch: 0.0199
204
- [2026-04-04 14:03:05] Step 3990: loss: 0.8577, grad_norm: 0.2070, learning_rate: 0.0000, epoch: 0.0199
205
- [2026-04-04 14:04:30] Step 4000: loss: 0.8141, grad_norm: 0.1709, learning_rate: 0.0000, epoch: 0.0200
206
- [2026-04-04 14:07:12] Step 4010: loss: 0.8507, grad_norm: 0.1602, learning_rate: 0.0000, epoch: 0.0200
207
- [2026-04-04 14:08:37] Step 4020: loss: 0.8619, grad_norm: 0.1729, learning_rate: 0.0000, epoch: 0.0201
208
- [2026-04-04 14:10:01] Step 4030: loss: 0.8497, grad_norm: 0.1514, learning_rate: 0.0000, epoch: 0.0202
209
- [2026-04-04 14:11:26] Step 4040: loss: 0.8421, grad_norm: 0.1729, learning_rate: 0.0000, epoch: 0.0202
210
- [2026-04-04 14:12:50] Step 4050: loss: 0.8204, grad_norm: 0.2129, learning_rate: 0.0000, epoch: 0.0203
211
- [2026-04-04 14:14:14] Step 4060: loss: 0.8450, grad_norm: 0.2236, learning_rate: 0.0000, epoch: 0.0203
212
- [2026-04-04 14:15:39] Step 4070: loss: 0.7626, grad_norm: 0.1138, learning_rate: 0.0000, epoch: 0.0204
213
- [2026-04-04 14:17:03] Step 4080: loss: 0.6547, grad_norm: 0.1084, learning_rate: 0.0000, epoch: 0.0204
214
- [2026-04-04 14:18:28] Step 4090: loss: 0.9119, grad_norm: 0.2070, learning_rate: 0.0000, epoch: 0.0204
215
- [2026-04-04 14:19:52] Step 4100: loss: 0.7530, grad_norm: 0.2500, learning_rate: 0.0000, epoch: 0.0205
216
- [2026-04-04 14:21:17] Step 4110: loss: 0.8678, grad_norm: 0.1592, learning_rate: 0.0000, epoch: 0.0205
217
- [2026-04-04 14:22:41] Step 4120: loss: 0.8312, grad_norm: 0.1982, learning_rate: 0.0000, epoch: 0.0206
218
- [2026-04-04 14:24:06] Step 4130: loss: 0.8719, grad_norm: 0.4414, learning_rate: 0.0000, epoch: 0.0207
219
- [2026-04-04 14:25:30] Step 4140: loss: 0.8141, grad_norm: 0.2598, learning_rate: 0.0000, epoch: 0.0207
220
- [2026-04-04 14:26:55] Step 4150: loss: 0.8780, grad_norm: 0.2324, learning_rate: 0.0000, epoch: 0.0208
221
- [2026-04-04 14:28:19] Step 4160: loss: 0.8864, grad_norm: 0.1514, learning_rate: 0.0000, epoch: 0.0208
222
- [2026-04-04 14:29:44] Step 4170: loss: 0.9051, grad_norm: 0.1611, learning_rate: 0.0000, epoch: 0.0209
223
- [2026-04-04 14:31:08] Step 4180: loss: 0.8238, grad_norm: 0.2676, learning_rate: 0.0000, epoch: 0.0209
224
- [2026-04-04 14:32:32] Step 4190: loss: 0.9161, grad_norm: 0.1904, learning_rate: 0.0000, epoch: 0.0209
225
- [2026-04-04 14:33:56] Step 4200: loss: 0.9077, grad_norm: 0.2158, learning_rate: 0.0000, epoch: 0.0210
226
- [2026-04-04 14:35:21] Step 4210: loss: 0.8494, grad_norm: 0.1465, learning_rate: 0.0000, epoch: 0.0210
227
- [2026-04-04 14:36:45] Step 4220: loss: 0.9484, grad_norm: 0.1855, learning_rate: 0.0000, epoch: 0.0211
228
- [2026-04-04 14:38:09] Step 4230: loss: 0.8625, grad_norm: 0.2236, learning_rate: 0.0000, epoch: 0.0211
229
- [2026-04-04 14:39:33] Step 4240: loss: 0.8828, grad_norm: 0.2031, learning_rate: 0.0000, epoch: 0.0212
230
- [2026-04-04 14:40:58] Step 4250: loss: 0.8607, grad_norm: 0.1572, learning_rate: 0.0000, epoch: 0.0213
231
- [2026-04-04 14:42:22] Step 4260: loss: 0.9015, grad_norm: 0.2129, learning_rate: 0.0000, epoch: 0.0213
232
- [2026-04-04 14:43:46] Step 4270: loss: 0.9145, grad_norm: 0.1748, learning_rate: 0.0000, epoch: 0.0214
233
- [2026-04-04 14:45:10] Step 4280: loss: 0.8502, grad_norm: 0.1914, learning_rate: 0.0000, epoch: 0.0214
234
- [2026-04-04 14:46:34] Step 4290: loss: 0.8569, grad_norm: 0.2432, learning_rate: 0.0000, epoch: 0.0215
235
- [2026-04-04 14:47:58] Step 4300: loss: 0.8992, grad_norm: 0.1865, learning_rate: 0.0000, epoch: 0.0215
236
- [2026-04-04 14:49:23] Step 4310: loss: 0.9081, grad_norm: 0.1641, learning_rate: 0.0000, epoch: 0.0215
237
- [2026-04-04 14:50:47] Step 4320: loss: 0.9125, grad_norm: 0.2012, learning_rate: 0.0000, epoch: 0.0216
238
- [2026-04-04 14:52:11] Step 4330: loss: 0.8991, grad_norm: 0.2295, learning_rate: 0.0000, epoch: 0.0216
239
- [2026-04-04 14:53:35] Step 4340: loss: 0.9418, grad_norm: 0.1611, learning_rate: 0.0000, epoch: 0.0217
240
- [2026-04-04 14:55:00] Step 4350: loss: 1.0309, grad_norm: 0.6445, learning_rate: 0.0000, epoch: 0.0217
241
- [2026-04-04 14:56:24] Step 4360: loss: 0.8411, grad_norm: 0.1279, learning_rate: 0.0000, epoch: 0.0218
242
- [2026-04-04 14:57:49] Step 4370: loss: 0.9025, grad_norm: 0.2930, learning_rate: 0.0000, epoch: 0.0219
243
- [2026-04-04 14:59:13] Step 4380: loss: 0.8707, grad_norm: 0.1768, learning_rate: 0.0000, epoch: 0.0219
244
- [2026-04-04 15:00:37] Step 4390: loss: 0.8823, grad_norm: 0.2197, learning_rate: 0.0000, epoch: 0.0220
245
- [2026-04-04 15:02:02] Step 4400: loss: 0.9306, grad_norm: 0.1885, learning_rate: 0.0000, epoch: 0.0220
246
- [2026-04-04 15:03:26] Step 4410: loss: 0.8781, grad_norm: 0.2041, learning_rate: 0.0000, epoch: 0.0221
247
- [2026-04-04 15:04:50] Step 4420: loss: 0.7951, grad_norm: 0.1895, learning_rate: 0.0000, epoch: 0.0221
248
- [2026-04-04 15:06:14] Step 4430: loss: 0.8375, grad_norm: 0.2324, learning_rate: 0.0000, epoch: 0.0221
249
- [2026-04-04 15:07:38] Step 4440: loss: 0.8831, grad_norm: 0.2949, learning_rate: 0.0000, epoch: 0.0222
250
- [2026-04-04 15:09:02] Step 4450: loss: 0.8990, grad_norm: 0.2275, learning_rate: 0.0000, epoch: 0.0222
251
- [2026-04-04 15:10:26] Step 4460: loss: 0.8656, grad_norm: 0.1650, learning_rate: 0.0000, epoch: 0.0223
252
- [2026-04-04 15:11:51] Step 4470: loss: 0.8416, grad_norm: 0.1777, learning_rate: 0.0000, epoch: 0.0223
253
- [2026-04-04 15:13:15] Step 4480: loss: 0.9821, grad_norm: 0.1807, learning_rate: 0.0000, epoch: 0.0224
254
- [2026-04-04 15:14:39] Step 4490: loss: 0.8345, grad_norm: 0.2490, learning_rate: 0.0000, epoch: 0.0225
255
- [2026-04-04 15:16:03] Step 4500: loss: 0.7619, grad_norm: 0.2246, learning_rate: 0.0000, epoch: 0.0225
256
- [2026-04-04 15:17:27] Step 4510: loss: 0.9103, grad_norm: 0.1846, learning_rate: 0.0000, epoch: 0.0226
257
- [2026-04-04 15:18:51] Step 4520: loss: 0.8421, grad_norm: 0.2129, learning_rate: 0.0000, epoch: 0.0226
258
- [2026-04-04 15:20:16] Step 4530: loss: 0.9850, grad_norm: 0.1963, learning_rate: 0.0000, epoch: 0.0226
259
- [2026-04-04 15:21:40] Step 4540: loss: 0.8908, grad_norm: 0.2139, learning_rate: 0.0000, epoch: 0.0227
260
- [2026-04-04 15:23:04] Step 4550: loss: 0.8702, grad_norm: 0.2061, learning_rate: 0.0000, epoch: 0.0227
261
- [2026-04-04 15:24:28] Step 4560: loss: 0.9466, grad_norm: 0.1914, learning_rate: 0.0000, epoch: 0.0228
262
- [2026-04-04 15:25:53] Step 4570: loss: 0.8640, grad_norm: 0.1709, learning_rate: 0.0000, epoch: 0.0228
263
- [2026-04-04 15:27:17] Step 4580: loss: 1.0906, grad_norm: 0.2451, learning_rate: 0.0000, epoch: 0.0229
264
- [2026-04-04 15:28:41] Step 4590: loss: 0.8248, grad_norm: 0.1689, learning_rate: 0.0000, epoch: 0.0230
265
- [2026-04-04 15:30:05] Step 4600: loss: 0.9053, grad_norm: 0.4121, learning_rate: 0.0000, epoch: 0.0230
266
- [2026-04-04 15:31:30] Step 4610: loss: 0.9402, grad_norm: 0.2158, learning_rate: 0.0000, epoch: 0.0231
267
- [2026-04-04 15:32:54] Step 4620: loss: 0.9434, grad_norm: 0.1650, learning_rate: 0.0000, epoch: 0.0231
268
- [2026-04-04 15:34:18] Step 4630: loss: 0.9944, grad_norm: 0.1953, learning_rate: 0.0000, epoch: 0.0232
269
- [2026-04-04 15:35:42] Step 4640: loss: 0.9170, grad_norm: 0.2031, learning_rate: 0.0000, epoch: 0.0232
270
- [2026-04-04 15:37:06] Step 4650: loss: 0.9447, grad_norm: 888.0000, learning_rate: 0.0000, epoch: 0.0232
271
- [2026-04-04 15:38:30] Step 4660: loss: 0.8551, grad_norm: 0.2402, learning_rate: 0.0000, epoch: 0.0233
272
- [2026-04-04 15:39:54] Step 4670: loss: 0.8951, grad_norm: 0.2119, learning_rate: 0.0000, epoch: 0.0233
273
- [2026-04-04 15:41:19] Step 4680: loss: 0.8970, grad_norm: 0.1885, learning_rate: 0.0000, epoch: 0.0234
274
- [2026-04-04 15:42:43] Step 4690: loss: 1.4014, grad_norm: 0.2158, learning_rate: 0.0000, epoch: 0.0234
275
- [2026-04-04 15:44:07] Step 4700: loss: 0.8219, grad_norm: 0.1777, learning_rate: 0.0000, epoch: 0.0235
276
- [2026-04-04 15:45:31] Step 4710: loss: 0.9223, grad_norm: 0.1982, learning_rate: 0.0000, epoch: 0.0236
277
- [2026-04-04 15:46:55] Step 4720: loss: 0.8339, grad_norm: 0.2715, learning_rate: 0.0000, epoch: 0.0236
278
- [2026-04-04 15:48:19] Step 4730: loss: 0.8905, grad_norm: 0.2305, learning_rate: 0.0000, epoch: 0.0237
279
- [2026-04-04 15:49:44] Step 4740: loss: 0.8911, grad_norm: 0.3223, learning_rate: 0.0000, epoch: 0.0237
280
- [2026-04-04 15:51:07] Step 4750: loss: 0.8771, grad_norm: 0.2910, learning_rate: 0.0000, epoch: 0.0238
281
- [2026-04-04 15:52:31] Step 4760: loss: 0.9284, grad_norm: 0.1641, learning_rate: 0.0000, epoch: 0.0238
282
- [2026-04-04 15:53:55] Step 4770: loss: 0.9055, grad_norm: 0.2451, learning_rate: 0.0000, epoch: 0.0238
283
- [2026-04-04 15:55:20] Step 4780: loss: 0.9106, grad_norm: 0.3027, learning_rate: 0.0000, epoch: 0.0239
284
- [2026-04-04 15:56:44] Step 4790: loss: 0.8667, grad_norm: 0.1611, learning_rate: 0.0000, epoch: 0.0239
285
- [2026-04-04 15:58:09] Step 4800: loss: 0.9344, grad_norm: 0.1729, learning_rate: 0.0000, epoch: 0.0240
286
- [2026-04-04 15:59:33] Step 4810: loss: 0.9068, grad_norm: 0.1748, learning_rate: 0.0000, epoch: 0.0240
287
- [2026-04-04 16:00:57] Step 4820: loss: 0.9075, grad_norm: 0.1465, learning_rate: 0.0000, epoch: 0.0241
288
- [2026-04-04 16:02:22] Step 4830: loss: 0.8334, grad_norm: 0.1621, learning_rate: 0.0000, epoch: 0.0242
289
- [2026-04-04 16:03:46] Step 4840: loss: 0.8115, grad_norm: 0.4375, learning_rate: 0.0000, epoch: 0.0242
290
- [2026-04-04 16:05:10] Step 4850: loss: 0.8499, grad_norm: 0.2188, learning_rate: 0.0000, epoch: 0.0243
291
- [2026-04-04 16:06:35] Step 4860: loss: 0.9564, grad_norm: 0.1826, learning_rate: 0.0000, epoch: 0.0243
292
- [2026-04-04 16:07:59] Step 4870: loss: 0.9589, grad_norm: 0.1670, learning_rate: 0.0000, epoch: 0.0244
293
- [2026-04-04 16:09:23] Step 4880: loss: 0.9117, grad_norm: 0.2178, learning_rate: 0.0000, epoch: 0.0244
294
- [2026-04-04 16:10:47] Step 4890: loss: 0.7845, grad_norm: 0.1338, learning_rate: 0.0000, epoch: 0.0244
295
- [2026-04-04 16:12:11] Step 4900: loss: 0.8921, grad_norm: 0.1738, learning_rate: 0.0000, epoch: 0.0245
296
- [2026-04-04 16:13:35] Step 4910: loss: 0.8833, grad_norm: 0.1396, learning_rate: 0.0000, epoch: 0.0245
297
- [2026-04-04 16:14:59] Step 4920: loss: 0.7464, grad_norm: 0.1934, learning_rate: 0.0000, epoch: 0.0246
298
- [2026-04-04 16:16:23] Step 4930: loss: 0.8945, grad_norm: 0.1680, learning_rate: 0.0000, epoch: 0.0246
299
- [2026-04-04 16:17:47] Step 4940: loss: 0.8504, grad_norm: 0.1836, learning_rate: 0.0000, epoch: 0.0247
300
- [2026-04-04 16:19:11] Step 4950: loss: 0.7694, grad_norm: 0.1875, learning_rate: 0.0000, epoch: 0.0248
301
- [2026-04-04 16:20:35] Step 4960: loss: 0.7889, grad_norm: 0.1270, learning_rate: 0.0000, epoch: 0.0248
302
- [2026-04-04 16:21:59] Step 4970: loss: 0.9033, grad_norm: 0.1777, learning_rate: 0.0000, epoch: 0.0249
303
- [2026-04-04 16:23:23] Step 4980: loss: 0.8958, grad_norm: 0.1758, learning_rate: 0.0000, epoch: 0.0249
304
- [2026-04-04 16:24:48] Step 4990: loss: 0.8875, grad_norm: 0.1924, learning_rate: 0.0000, epoch: 0.0249
305
- [2026-04-04 16:26:12] Step 5000: loss: 0.8543, grad_norm: 0.2471, learning_rate: 0.0000, epoch: 0.0250
306
- [2026-04-04 16:27:36] Step 5010: loss: 0.8250, grad_norm: 0.1826, learning_rate: 0.0000, epoch: 0.0250
307
- [2026-04-04 16:29:01] Step 5020: loss: 0.8954, grad_norm: 0.2100, learning_rate: 0.0000, epoch: 0.0251
308
- [2026-04-04 16:30:26] Step 5030: loss: 0.7521, grad_norm: 0.1836, learning_rate: 0.0000, epoch: 0.0251
309
- [2026-04-04 16:31:51] Step 5040: loss: 0.7933, grad_norm: 0.1377, learning_rate: 0.0000, epoch: 0.0252
310
- [2026-04-04 16:33:15] Step 5050: loss: 0.8361, grad_norm: 0.1348, learning_rate: 0.0000, epoch: 0.0253
311
- [2026-04-04 16:34:40] Step 5060: loss: 0.8986, grad_norm: 0.1514, learning_rate: 0.0000, epoch: 0.0253
312
- [2026-04-04 16:36:06] Step 5070: loss: 0.8269, grad_norm: 0.2285, learning_rate: 0.0000, epoch: 0.0254
313
- [2026-04-04 16:37:30] Step 5080: loss: 0.9033, grad_norm: 0.1572, learning_rate: 0.0000, epoch: 0.0254
314
- [2026-04-04 16:38:56] Step 5090: loss: 0.8741, grad_norm: 0.1484, learning_rate: 0.0000, epoch: 0.0255
315
- [2026-04-04 16:40:21] Step 5100: loss: 0.7754, grad_norm: 0.1611, learning_rate: 0.0000, epoch: 0.0255
316
- [2026-04-04 16:41:45] Step 5110: loss: 0.7053, grad_norm: 0.1445, learning_rate: 0.0000, epoch: 0.0255
317
- [2026-04-04 16:43:10] Step 5120: loss: 0.8493, grad_norm: 0.1514, learning_rate: 0.0000, epoch: 0.0256
318
- [2026-04-04 16:44:34] Step 5130: loss: 0.8398, grad_norm: 0.1494, learning_rate: 0.0000, epoch: 0.0256
319
- [2026-04-04 16:45:59] Step 5140: loss: 0.8019, grad_norm: 0.1650, learning_rate: 0.0000, epoch: 0.0257
320
- [2026-04-04 16:47:23] Step 5150: loss: 0.7548, grad_norm: 0.2012, learning_rate: 0.0000, epoch: 0.0257
321
- [2026-04-04 16:48:48] Step 5160: loss: 0.8675, grad_norm: 0.1621, learning_rate: 0.0000, epoch: 0.0258
322
- [2026-04-04 16:50:13] Step 5170: loss: 0.7951, grad_norm: 0.1934, learning_rate: 0.0000, epoch: 0.0259
323
- [2026-04-04 16:51:39] Step 5180: loss: 0.9608, grad_norm: 0.2051, learning_rate: 0.0000, epoch: 0.0259
324
- [2026-04-04 16:53:03] Step 5190: loss: 0.9374, grad_norm: 0.2090, learning_rate: 0.0000, epoch: 0.0260
325
- [2026-04-04 16:54:28] Step 5200: loss: 0.8681, grad_norm: 0.1270, learning_rate: 0.0000, epoch: 0.0260
326
- [2026-04-04 16:55:52] Step 5210: loss: 0.8505, grad_norm: 0.1699, learning_rate: 0.0000, epoch: 0.0261
327
- [2026-04-04 16:57:17] Step 5220: loss: 0.8585, grad_norm: 0.1270, learning_rate: 0.0000, epoch: 0.0261
328
- [2026-04-04 16:58:42] Step 5230: loss: 0.7865, grad_norm: 0.2500, learning_rate: 0.0000, epoch: 0.0261
329
- [2026-04-04 17:00:07] Step 5240: loss: 0.8359, grad_norm: 0.1680, learning_rate: 0.0000, epoch: 0.0262
330
- [2026-04-04 17:01:32] Step 5250: loss: 0.9111, grad_norm: 0.2471, learning_rate: 0.0000, epoch: 0.0262
331
- [2026-04-04 17:02:56] Step 5260: loss: 0.8239, grad_norm: 0.1641, learning_rate: 0.0000, epoch: 0.0263
332
- [2026-04-04 17:04:21] Step 5270: loss: 0.9209, grad_norm: 0.1797, learning_rate: 0.0000, epoch: 0.0263
333
- [2026-04-04 17:05:46] Step 5280: loss: 0.9036, grad_norm: 0.1895, learning_rate: 0.0000, epoch: 0.0264
334
- [2026-04-04 17:07:10] Step 5290: loss: 0.8678, grad_norm: 0.2139, learning_rate: 0.0000, epoch: 0.0265
335
- [2026-04-04 17:08:34] Step 5300: loss: 0.8450, grad_norm: 0.3770, learning_rate: 0.0000, epoch: 0.0265
336
- [2026-04-04 17:09:58] Step 5310: loss: 0.8964, grad_norm: 0.1777, learning_rate: 0.0000, epoch: 0.0266
337
- [2026-04-04 17:11:23] Step 5320: loss: 0.8130, grad_norm: 0.2119, learning_rate: 0.0000, epoch: 0.0266
338
- [2026-04-04 17:12:47] Step 5330: loss: 0.8837, grad_norm: 0.2197, learning_rate: 0.0000, epoch: 0.0267
339
- [2026-04-04 17:14:12] Step 5340: loss: 0.9022, grad_norm: 0.1465, learning_rate: 0.0000, epoch: 0.0267
340
- [2026-04-04 17:15:36] Step 5350: loss: 0.9066, grad_norm: 0.2070, learning_rate: 0.0000, epoch: 0.0267
341
- [2026-04-04 17:17:01] Step 5360: loss: 0.8575, grad_norm: 0.1660, learning_rate: 0.0000, epoch: 0.0268
342
- [2026-04-04 17:18:25] Step 5370: loss: 1.0600, grad_norm: 0.3203, learning_rate: 0.0000, epoch: 0.0268
343
- [2026-04-04 17:19:50] Step 5380: loss: 0.8405, grad_norm: 0.2393, learning_rate: 0.0000, epoch: 0.0269
344
- [2026-04-04 17:21:15] Step 5390: loss: 0.8278, grad_norm: 0.1982, learning_rate: 0.0000, epoch: 0.0270
345
- [2026-04-04 17:22:39] Step 5400: loss: 0.9193, grad_norm: 0.1982, learning_rate: 0.0000, epoch: 0.0270
346
- [2026-04-04 17:24:03] Step 5410: loss: 0.9441, grad_norm: 0.3066, learning_rate: 0.0000, epoch: 0.0271
347
- [2026-04-04 17:25:28] Step 5420: loss: 0.8435, grad_norm: 0.1914, learning_rate: 0.0000, epoch: 0.0271
348
- [2026-04-04 17:26:53] Step 5430: loss: 0.8175, grad_norm: 0.1523, learning_rate: 0.0000, epoch: 0.0272
349
- [2026-04-04 17:28:17] Step 5440: loss: 0.8007, grad_norm: 0.2695, learning_rate: 0.0000, epoch: 0.0272
350
- [2026-04-04 17:29:41] Step 5450: loss: 0.9872, grad_norm: 0.2461, learning_rate: 0.0000, epoch: 0.0272
351
- [2026-04-04 17:31:06] Step 5460: loss: 0.8669, grad_norm: 0.2080, learning_rate: 0.0000, epoch: 0.0273
352
- [2026-04-04 17:32:30] Step 5470: loss: 0.8589, grad_norm: 0.3223, learning_rate: 0.0000, epoch: 0.0273
353
- [2026-04-04 17:33:54] Step 5480: loss: 0.9051, grad_norm: 0.1758, learning_rate: 0.0000, epoch: 0.0274
354
- [2026-04-04 17:35:19] Step 5490: loss: 0.9042, grad_norm: 0.2539, learning_rate: 0.0000, epoch: 0.0274
355
- [2026-04-04 17:36:44] Step 5500: loss: 0.8508, grad_norm: 0.1367, learning_rate: 0.0000, epoch: 0.0275
356
- [2026-04-04 17:38:08] Step 5510: loss: 0.9220, grad_norm: 0.1475, learning_rate: 0.0000, epoch: 0.0276
357
- [2026-04-04 17:39:33] Step 5520: loss: 0.9046, grad_norm: 0.5312, learning_rate: 0.0000, epoch: 0.0276
358
- [2026-04-04 17:40:58] Step 5530: loss: 0.8969, grad_norm: 0.2002, learning_rate: 0.0000, epoch: 0.0277
359
- [2026-04-04 17:42:22] Step 5540: loss: 0.9212, grad_norm: 0.2910, learning_rate: 0.0000, epoch: 0.0277
360
- [2026-04-04 17:43:46] Step 5550: loss: 0.8605, grad_norm: 0.2930, learning_rate: 0.0000, epoch: 0.0278
361
- [2026-04-04 17:45:11] Step 5560: loss: 0.9055, grad_norm: 0.2676, learning_rate: 0.0000, epoch: 0.0278
362
- [2026-04-04 17:46:35] Step 5570: loss: 0.9778, grad_norm: 0.1689, learning_rate: 0.0000, epoch: 0.0278
363
- [2026-04-04 17:48:00] Step 5580: loss: 0.9480, grad_norm: 0.2207, learning_rate: 0.0000, epoch: 0.0279
364
- [2026-04-04 17:49:24] Step 5590: loss: 0.9285, grad_norm: 0.2354, learning_rate: 0.0000, epoch: 0.0279
365
- [2026-04-04 17:50:48] Step 5600: loss: 0.9215, grad_norm: 0.2637, learning_rate: 0.0000, epoch: 0.0280
366
- [2026-04-04 17:52:12] Step 5610: loss: 0.8681, grad_norm: 0.1973, learning_rate: 0.0000, epoch: 0.0280
367
- [2026-04-04 17:53:37] Step 5620: loss: 0.9485, grad_norm: 0.2393, learning_rate: 0.0000, epoch: 0.0281
368
- [2026-04-04 17:55:01] Step 5630: loss: 0.9471, grad_norm: 5.0938, learning_rate: 0.0000, epoch: 0.0282
369
- [2026-04-04 17:56:26] Step 5640: loss: 0.9299, grad_norm: 0.1602, learning_rate: 0.0000, epoch: 0.0282
370
- [2026-04-04 17:57:50] Step 5650: loss: 0.9612, grad_norm: 0.6289, learning_rate: 0.0000, epoch: 0.0283
371
- [2026-04-04 17:59:15] Step 5660: loss: 0.8904, grad_norm: 0.3438, learning_rate: 0.0000, epoch: 0.0283
372
- [2026-04-04 18:00:39] Step 5670: loss: 0.7939, grad_norm: 0.2119, learning_rate: 0.0000, epoch: 0.0284
373
- [2026-04-04 18:02:03] Step 5680: loss: 0.8262, grad_norm: 0.2168, learning_rate: 0.0000, epoch: 0.0284
374
- [2026-04-04 18:03:28] Step 5690: loss: 0.8845, grad_norm: 0.2412, learning_rate: 0.0000, epoch: 0.0284
375
- [2026-04-04 18:04:52] Step 5700: loss: 0.9226, grad_norm: 0.1738, learning_rate: 0.0000, epoch: 0.0285
376
- [2026-04-04 18:06:17] Step 5710: loss: 0.8747, grad_norm: 0.3926, learning_rate: 0.0000, epoch: 0.0285
377
- [2026-04-04 18:07:42] Step 5720: loss: 0.8984, grad_norm: 0.1768, learning_rate: 0.0000, epoch: 0.0286
378
- [2026-04-04 18:09:06] Step 5730: loss: 0.8789, grad_norm: 0.1709, learning_rate: 0.0000, epoch: 0.0286
379
- [2026-04-04 18:10:29] Step 5740: loss: 1.0040, grad_norm: 0.5234, learning_rate: 0.0000, epoch: 0.0287
380
- [2026-04-04 18:11:54] Step 5750: loss: 0.8954, grad_norm: 0.2109, learning_rate: 0.0000, epoch: 0.0288
381
- [2026-04-04 18:13:18] Step 5760: loss: 0.8071, grad_norm: 0.1533, learning_rate: 0.0000, epoch: 0.0288
382
- [2026-04-04 18:14:42] Step 5770: loss: 0.9445, grad_norm: 0.1543, learning_rate: 0.0000, epoch: 0.0289
383
- [2026-04-04 18:16:07] Step 5780: loss: 0.8859, grad_norm: 1592.0000, learning_rate: 0.0000, epoch: 0.0289
384
- [2026-04-04 18:17:31] Step 5790: loss: 0.9482, grad_norm: 0.1836, learning_rate: 0.0000, epoch: 0.0290
385
- [2026-04-04 18:18:56] Step 5800: loss: 0.9173, grad_norm: 0.2432, learning_rate: 0.0000, epoch: 0.0290
386
- [2026-04-04 18:20:20] Step 5810: loss: 0.9080, grad_norm: 0.1533, learning_rate: 0.0000, epoch: 0.0290
387
- [2026-04-04 18:21:45] Step 5820: loss: 0.7803, grad_norm: 0.1875, learning_rate: 0.0000, epoch: 0.0291
388
- [2026-04-04 18:23:09] Step 5830: loss: 0.8363, grad_norm: 0.1719, learning_rate: 0.0000, epoch: 0.0291
389
- [2026-04-04 18:24:33] Step 5840: loss: 0.9606, grad_norm: 0.3848, learning_rate: 0.0000, epoch: 0.0292
390
- [2026-04-04 18:25:57] Step 5850: loss: 0.8659, grad_norm: 0.2227, learning_rate: 0.0000, epoch: 0.0293
391
- [2026-04-04 18:27:21] Step 5860: loss: 0.7850, grad_norm: 0.2256, learning_rate: 0.0000, epoch: 0.0293
392
- [2026-04-04 18:28:46] Step 5870: loss: 0.8648, grad_norm: 0.2158, learning_rate: 0.0000, epoch: 0.0294
393
- [2026-04-04 18:30:10] Step 5880: loss: 0.9564, grad_norm: 0.2461, learning_rate: 0.0000, epoch: 0.0294
394
- [2026-04-04 18:31:34] Step 5890: loss: 0.9538, grad_norm: 0.1855, learning_rate: 0.0000, epoch: 0.0295
395
- [2026-04-04 18:32:59] Step 5900: loss: 0.8727, grad_norm: 0.1875, learning_rate: 0.0000, epoch: 0.0295
396
- [2026-04-04 18:34:23] Step 5910: loss: 0.9268, grad_norm: 3.0781, learning_rate: 0.0000, epoch: 0.0295
397
- [2026-04-04 18:35:47] Step 5920: loss: 0.8603, grad_norm: 0.1797, learning_rate: 0.0000, epoch: 0.0296
398
- [2026-04-04 18:37:11] Step 5930: loss: 0.9185, grad_norm: 0.3555, learning_rate: 0.0000, epoch: 0.0296
399
- [2026-04-04 18:38:36] Step 5940: loss: 0.8986, grad_norm: 0.1562, learning_rate: 0.0000, epoch: 0.0297
400
- [2026-04-04 18:40:00] Step 5950: loss: 0.9599, grad_norm: 0.1826, learning_rate: 0.0000, epoch: 0.0297
401
- [2026-04-04 18:41:24] Step 5960: loss: 0.8563, grad_norm: 0.2002, learning_rate: 0.0000, epoch: 0.0298
402
- [2026-04-04 18:42:49] Step 5970: loss: 0.8872, grad_norm: 0.4316, learning_rate: 0.0000, epoch: 0.0299
403
- [2026-04-04 18:44:13] Step 5980: loss: 0.7684, grad_norm: 0.1934, learning_rate: 0.0000, epoch: 0.0299
404
- [2026-04-04 18:45:37] Step 5990: loss: 0.8413, grad_norm: 0.2578, learning_rate: 0.0000, epoch: 0.0300
405
- [2026-04-04 18:47:01] Step 6000: loss: 0.9206, grad_norm: 0.2090, learning_rate: 0.0000, epoch: 0.0300
406
- [2026-04-04 18:50:59] Step 6010: loss: 0.8848, grad_norm: 0.1934, learning_rate: 0.0000, epoch: 0.0301
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
 
2
  ==================================================
3
+ Training started at: 2026-04-05 10:32:25
4
  ==================================================
5
 
6
+ [2026-04-05 10:34:27] Step 1: loss: 5.374693, grad_norm: 704.000000, learning_rate: 0.000000, epoch: 0.000005
7
+ [2026-04-05 10:35:42] Step 10: loss: 6.319018, grad_norm: 44.250000, learning_rate: 0.000001, epoch: 0.000051
8
+ [2026-04-05 10:37:07] Step 20: loss: 4.786968, grad_norm: 0.894531, learning_rate: 0.000002, epoch: 0.000101
9
+ [2026-04-05 10:38:33] Step 30: loss: 4.273304, grad_norm: 1.117188, learning_rate: 0.000003, epoch: 0.000152
10
+ [2026-04-05 10:39:58] Step 40: loss: 4.699950, grad_norm: 0.996094, learning_rate: 0.000004, epoch: 0.000202
11
+ [2026-04-05 10:41:24] Step 50: loss: 3.808567, grad_norm: 0.429688, learning_rate: 0.000005, epoch: 0.000253
12
+ [2026-04-05 10:42:49] Step 60: loss: 4.120002, grad_norm: 0.333984, learning_rate: 0.000006, epoch: 0.000303
13
+ [2026-04-05 10:44:15] Step 70: loss: 3.806800, grad_norm: 0.679688, learning_rate: 0.000007, epoch: 0.000354
14
+ [2026-04-05 10:45:40] Step 80: loss: 4.653148, grad_norm: 1.046875, learning_rate: 0.000008, epoch: 0.000404
15
+ [2026-04-05 10:47:05] Step 90: loss: 3.986430, grad_norm: 0.746094, learning_rate: 0.000009, epoch: 0.000455
16
+ [2026-04-05 10:48:31] Step 100: loss: 3.177014, grad_norm: 0.972656, learning_rate: 0.000010, epoch: 0.000505
17
+ [2026-04-05 10:49:58] Step 110: loss: 2.341252, grad_norm: 0.636719, learning_rate: 0.000011, epoch: 0.000556
18
+ [2026-04-05 10:51:23] Step 120: loss: 3.686772, grad_norm: 0.496094, learning_rate: 0.000012, epoch: 0.000606
19
+ [2026-04-05 10:52:48] Step 130: loss: 4.326537, grad_norm: 4.031250, learning_rate: 0.000013, epoch: 0.000657
20
+ [2026-04-05 10:54:14] Step 140: loss: 2.882456, grad_norm: 0.248047, learning_rate: 0.000014, epoch: 0.000707
21
+ [2026-04-05 10:55:39] Step 150: loss: 3.564154, grad_norm: 1.023438, learning_rate: 0.000015, epoch: 0.000758
22
+ [2026-04-05 10:57:05] Step 160: loss: 3.621474, grad_norm: 0.574219, learning_rate: 0.000016, epoch: 0.000808
23
+ [2026-04-05 10:58:31] Step 170: loss: 3.432347, grad_norm: 0.247070, learning_rate: 0.000017, epoch: 0.000859
24
+ [2026-04-05 10:59:55] Step 180: loss: 3.368018, grad_norm: 0.339844, learning_rate: 0.000018, epoch: 0.000909
25
+ [2026-04-05 11:01:21] Step 190: loss: 2.903378, grad_norm: 1.210938, learning_rate: 0.000019, epoch: 0.000960
26
+ [2026-04-05 11:02:47] Step 200: loss: 2.334621, grad_norm: 0.285156, learning_rate: 0.000020, epoch: 0.001010
27
+ [2026-04-05 11:04:11] Step 210: loss: 3.505114, grad_norm: 0.494141, learning_rate: 0.000021, epoch: 0.001061
28
+ [2026-04-05 11:05:36] Step 220: loss: 1.382823, grad_norm: 0.320312, learning_rate: 0.000022, epoch: 0.001111
29
+ [2026-04-05 11:07:02] Step 230: loss: 0.917196, grad_norm: 0.371094, learning_rate: 0.000023, epoch: 0.001162
30
+ [2026-04-05 11:08:29] Step 240: loss: 0.975546, grad_norm: 0.267578, learning_rate: 0.000024, epoch: 0.001212
31
+ [2026-04-05 11:09:55] Step 250: loss: 0.629887, grad_norm: 0.228516, learning_rate: 0.000025, epoch: 0.001263
32
+ [2026-04-05 11:11:22] Step 260: loss: 0.699167, grad_norm: 1.085938, learning_rate: 0.000026, epoch: 0.001313
33
+ [2026-04-05 11:12:49] Step 270: loss: 0.739236, grad_norm: 0.455078, learning_rate: 0.000027, epoch: 0.001364
34
+ [2026-04-05 11:14:15] Step 280: loss: 0.738538, grad_norm: 0.174805, learning_rate: 0.000028, epoch: 0.001414
35
+ [2026-04-05 11:15:40] Step 290: loss: 0.670122, grad_norm: 0.223633, learning_rate: 0.000029, epoch: 0.001465
36
+ [2026-04-05 11:17:06] Step 300: loss: 0.608715, grad_norm: 0.277344, learning_rate: 0.000030, epoch: 0.001515
37
+ [2026-04-05 11:18:31] Step 310: loss: 0.838763, grad_norm: 0.232422, learning_rate: 0.000031, epoch: 0.001566
38
+ [2026-04-05 11:19:57] Step 320: loss: 0.797982, grad_norm: 0.578125, learning_rate: 0.000032, epoch: 0.001616
39
+ [2026-04-05 11:21:24] Step 330: loss: 0.644051, grad_norm: 0.185547, learning_rate: 0.000033, epoch: 0.001667
40
+ [2026-04-05 11:22:49] Step 340: loss: 0.777247, grad_norm: 0.236328, learning_rate: 0.000034, epoch: 0.001717
41
+ [2026-04-05 11:24:16] Step 350: loss: 0.741875, grad_norm: 0.182617, learning_rate: 0.000035, epoch: 0.001768
42
+ [2026-04-05 11:25:42] Step 360: loss: 0.585453, grad_norm: 0.451172, learning_rate: 0.000036, epoch: 0.001818
43
+ [2026-04-05 11:27:08] Step 370: loss: 0.728102, grad_norm: 0.271484, learning_rate: 0.000037, epoch: 0.001869
44
+ [2026-04-05 11:28:35] Step 380: loss: 0.672046, grad_norm: 0.185547, learning_rate: 0.000038, epoch: 0.001919
45
+ [2026-04-05 11:30:01] Step 390: loss: 0.613392, grad_norm: 0.306641, learning_rate: 0.000039, epoch: 0.001970
46
+ [2026-04-05 11:31:26] Step 400: loss: 0.712388, grad_norm: 0.179688, learning_rate: 0.000040, epoch: 0.002020
47
+ [2026-04-05 11:32:53] Step 410: loss: 0.642185, grad_norm: 0.223633, learning_rate: 0.000041, epoch: 0.002071
48
+ [2026-04-05 11:34:18] Step 420: loss: 0.583045, grad_norm: 0.492188, learning_rate: 0.000042, epoch: 0.002121
49
+ [2026-04-05 11:35:44] Step 430: loss: 0.668336, grad_norm: 0.179688, learning_rate: 0.000043, epoch: 0.002172
50
+ [2026-04-05 11:37:10] Step 440: loss: 0.709393, grad_norm: 0.185547, learning_rate: 0.000044, epoch: 0.002222
51
+ [2026-04-05 11:38:36] Step 450: loss: 0.799581, grad_norm: 0.263672, learning_rate: 0.000045, epoch: 0.002273
52
+ [2026-04-05 11:40:02] Step 460: loss: 0.613966, grad_norm: 0.314453, learning_rate: 0.000046, epoch: 0.002323
53
+ [2026-04-05 11:41:28] Step 470: loss: 0.782370, grad_norm: 0.261719, learning_rate: 0.000047, epoch: 0.002374
54
+ [2026-04-05 11:42:54] Step 480: loss: 0.686781, grad_norm: 0.972656, learning_rate: 0.000048, epoch: 0.002424
55
+ [2026-04-05 11:44:20] Step 490: loss: 0.691988, grad_norm: 0.310547, learning_rate: 0.000049, epoch: 0.002475
56
+ [2026-04-05 11:45:45] Step 500: loss: 0.876984, grad_norm: 0.369141, learning_rate: 0.000050, epoch: 0.002525
57
+ [2026-04-05 11:47:11] Step 510: loss: 0.864422, grad_norm: 27.125000, learning_rate: 0.000050, epoch: 0.002576
58
+ [2026-04-05 11:48:37] Step 520: loss: 0.639947, grad_norm: 3.015625, learning_rate: 0.000050, epoch: 0.002626
59
+ [2026-04-05 11:50:03] Step 530: loss: 0.694945, grad_norm: 0.267578, learning_rate: 0.000050, epoch: 0.002677
60
+ [2026-04-05 11:51:29] Step 540: loss: 0.740541, grad_norm: 0.257812, learning_rate: 0.000050, epoch: 0.002727
61
+ [2026-04-05 11:52:55] Step 550: loss: 0.665531, grad_norm: 1.664062, learning_rate: 0.000050, epoch: 0.002778
62
+ [2026-04-05 11:54:22] Step 560: loss: 0.724073, grad_norm: 0.261719, learning_rate: 0.000050, epoch: 0.002828
63
+ [2026-04-05 11:55:48] Step 570: loss: 0.746542, grad_norm: 0.243164, learning_rate: 0.000050, epoch: 0.002879
64
+ [2026-04-05 11:57:15] Step 580: loss: 0.692906, grad_norm: 0.194336, learning_rate: 0.000050, epoch: 0.002929
65
+ [2026-04-05 11:58:41] Step 590: loss: 0.690486, grad_norm: 0.341797, learning_rate: 0.000050, epoch: 0.002980
66
+ [2026-04-05 12:00:07] Step 600: loss: 0.650022, grad_norm: 0.289062, learning_rate: 0.000050, epoch: 0.003030
67
+ [2026-04-05 12:01:34] Step 610: loss: 0.640746, grad_norm: 0.221680, learning_rate: 0.000050, epoch: 0.003081
68
+ [2026-04-05 12:03:00] Step 620: loss: 1.449164, grad_norm: 1.546875, learning_rate: 0.000050, epoch: 0.003131
69
+ [2026-04-05 12:04:26] Step 630: loss: 0.656220, grad_norm: 5.218750, learning_rate: 0.000050, epoch: 0.003182
70
+ [2026-04-05 12:05:53] Step 640: loss: 0.636382, grad_norm: 0.196289, learning_rate: 0.000050, epoch: 0.003232
71
+ [2026-04-05 12:07:19] Step 650: loss: 0.668934, grad_norm: 0.675781, learning_rate: 0.000050, epoch: 0.003283
72
+ [2026-04-05 12:08:45] Step 660: loss: 0.748885, grad_norm: 0.257812, learning_rate: 0.000050, epoch: 0.003333
73
+ [2026-04-05 12:10:11] Step 670: loss: 0.788645, grad_norm: 0.394531, learning_rate: 0.000050, epoch: 0.003384
74
+ [2026-04-05 12:11:36] Step 680: loss: 0.697456, grad_norm: 1.367188, learning_rate: 0.000050, epoch: 0.003434
75
+ [2026-04-05 12:13:03] Step 690: loss: 0.792172, grad_norm: 0.265625, learning_rate: 0.000050, epoch: 0.003485
76
+ [2026-04-05 12:14:29] Step 700: loss: 0.606292, grad_norm: 0.216797, learning_rate: 0.000050, epoch: 0.003535
77
+ [2026-04-05 12:15:55] Step 710: loss: 0.598090, grad_norm: 1.273438, learning_rate: 0.000050, epoch: 0.003586
78
+ [2026-04-05 12:17:21] Step 720: loss: 0.541651, grad_norm: 0.240234, learning_rate: 0.000050, epoch: 0.003636
79
+ [2026-04-05 12:18:47] Step 730: loss: 0.705516, grad_norm: 0.269531, learning_rate: 0.000050, epoch: 0.003687
80
+ [2026-04-05 12:20:13] Step 740: loss: 0.738221, grad_norm: 0.283203, learning_rate: 0.000050, epoch: 0.003737
81
+ [2026-04-05 12:21:39] Step 750: loss: 0.643487, grad_norm: 0.314453, learning_rate: 0.000050, epoch: 0.003788
82
+ [2026-04-05 12:23:06] Step 760: loss: 0.658763, grad_norm: 0.216797, learning_rate: 0.000050, epoch: 0.003838
83
+ [2026-04-05 12:24:31] Step 770: loss: 0.842584, grad_norm: 0.302734, learning_rate: 0.000050, epoch: 0.003889
84
+ [2026-04-05 12:25:56] Step 780: loss: 0.650289, grad_norm: 0.285156, learning_rate: 0.000050, epoch: 0.003939
85
+ [2026-04-05 12:27:22] Step 790: loss: 0.637068, grad_norm: 0.820312, learning_rate: 0.000050, epoch: 0.003990
86
+ [2026-04-05 12:28:48] Step 800: loss: 0.753985, grad_norm: 0.380859, learning_rate: 0.000050, epoch: 0.004040
87
+ [2026-04-05 12:30:14] Step 810: loss: 0.632987, grad_norm: 0.210938, learning_rate: 0.000050, epoch: 0.004091
88
+ [2026-04-05 12:31:41] Step 820: loss: 0.630292, grad_norm: 0.281250, learning_rate: 0.000050, epoch: 0.004141
89
+ [2026-04-05 12:33:05] Step 830: loss: 1.052068, grad_norm: 0.341797, learning_rate: 0.000050, epoch: 0.004192
90
+ [2026-04-05 12:34:31] Step 840: loss: 0.878007, grad_norm: 0.223633, learning_rate: 0.000050, epoch: 0.004242
91
+ [2026-04-05 12:35:57] Step 850: loss: 0.665595, grad_norm: 1.851562, learning_rate: 0.000050, epoch: 0.004293
92
+ [2026-04-05 12:37:24] Step 860: loss: 0.665583, grad_norm: 0.216797, learning_rate: 0.000050, epoch: 0.004343
93
+ [2026-04-05 12:38:51] Step 870: loss: 0.647318, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.004394
94
+ [2026-04-05 12:40:17] Step 880: loss: 0.662116, grad_norm: 0.166016, learning_rate: 0.000050, epoch: 0.004444
95
+ [2026-04-05 12:41:44] Step 890: loss: 0.642127, grad_norm: 0.190430, learning_rate: 0.000050, epoch: 0.004495
96
+ [2026-04-05 12:43:08] Step 900: loss: 0.716069, grad_norm: 0.253906, learning_rate: 0.000050, epoch: 0.004545
97
+ [2026-04-05 12:44:34] Step 910: loss: 0.595208, grad_norm: 0.687500, learning_rate: 0.000050, epoch: 0.004596
98
+ [2026-04-05 12:46:00] Step 920: loss: 0.696013, grad_norm: 0.226562, learning_rate: 0.000050, epoch: 0.004646
99
+ [2026-04-05 12:47:27] Step 930: loss: 0.703280, grad_norm: 0.498047, learning_rate: 0.000050, epoch: 0.004697
100
+ [2026-04-05 12:48:52] Step 940: loss: 0.597051, grad_norm: 0.330078, learning_rate: 0.000050, epoch: 0.004747
101
+ [2026-04-05 12:50:18] Step 950: loss: 0.667682, grad_norm: 0.179688, learning_rate: 0.000050, epoch: 0.004798
102
+ [2026-04-05 12:51:45] Step 960: loss: 0.640542, grad_norm: 0.153320, learning_rate: 0.000050, epoch: 0.004848
103
+ [2026-04-05 12:53:11] Step 970: loss: 0.671970, grad_norm: 0.427734, learning_rate: 0.000050, epoch: 0.004899
104
+ [2026-04-05 12:54:37] Step 980: loss: 0.693704, grad_norm: 0.229492, learning_rate: 0.000050, epoch: 0.004949
105
+ [2026-04-05 12:56:04] Step 990: loss: 0.625149, grad_norm: 0.636719, learning_rate: 0.000050, epoch: 0.005000
106
+ [2026-04-05 12:57:30] Step 1000: loss: 0.782581, grad_norm: 0.248047, learning_rate: 0.000050, epoch: 0.005051
107
+ [2026-04-05 12:58:56] Step 1010: loss: 0.668274, grad_norm: 0.181641, learning_rate: 0.000050, epoch: 0.005101
108
+ [2026-04-05 13:00:22] Step 1020: loss: 0.610654, grad_norm: 0.218750, learning_rate: 0.000050, epoch: 0.005152
109
+ [2026-04-05 13:01:49] Step 1030: loss: 0.662170, grad_norm: 0.150391, learning_rate: 0.000050, epoch: 0.005202
110
+ [2026-04-05 13:03:16] Step 1040: loss: 0.666967, grad_norm: 0.298828, learning_rate: 0.000050, epoch: 0.005253
111
+ [2026-04-05 13:04:41] Step 1050: loss: 0.680621, grad_norm: 0.190430, learning_rate: 0.000050, epoch: 0.005303
112
+ [2026-04-05 13:06:08] Step 1060: loss: 0.616938, grad_norm: 0.219727, learning_rate: 0.000050, epoch: 0.005354
113
+ [2026-04-05 13:07:34] Step 1070: loss: 0.619953, grad_norm: 0.142578, learning_rate: 0.000050, epoch: 0.005404
114
+ [2026-04-05 13:09:00] Step 1080: loss: 0.589728, grad_norm: 0.154297, learning_rate: 0.000050, epoch: 0.005455
115
+ [2026-04-05 13:10:27] Step 1090: loss: 0.573597, grad_norm: 0.179688, learning_rate: 0.000050, epoch: 0.005505
116
+ [2026-04-05 13:11:53] Step 1100: loss: 0.613491, grad_norm: 0.206055, learning_rate: 0.000050, epoch: 0.005556
117
+ [2026-04-05 13:13:18] Step 1110: loss: 0.460272, grad_norm: 0.214844, learning_rate: 0.000050, epoch: 0.005606
118
+ [2026-04-05 13:14:45] Step 1120: loss: 0.553796, grad_norm: 0.153320, learning_rate: 0.000050, epoch: 0.005657
119
+ [2026-04-05 13:16:11] Step 1130: loss: 0.724727, grad_norm: 0.198242, learning_rate: 0.000050, epoch: 0.005707
120
+ [2026-04-05 13:17:37] Step 1140: loss: 0.491195, grad_norm: 0.137695, learning_rate: 0.000050, epoch: 0.005758
121
+ [2026-04-05 13:19:01] Step 1150: loss: 0.529206, grad_norm: 0.191406, learning_rate: 0.000050, epoch: 0.005808
122
+ [2026-04-05 13:20:28] Step 1160: loss: 0.525564, grad_norm: 0.140625, learning_rate: 0.000050, epoch: 0.005859
123
+ [2026-04-05 13:21:55] Step 1170: loss: 0.576415, grad_norm: 0.273438, learning_rate: 0.000050, epoch: 0.005909
124
+ [2026-04-05 13:23:21] Step 1180: loss: 0.495870, grad_norm: 0.283203, learning_rate: 0.000050, epoch: 0.005960
125
+ [2026-04-05 13:24:47] Step 1190: loss: 0.691925, grad_norm: 0.213867, learning_rate: 0.000050, epoch: 0.006010
126
+ [2026-04-05 13:26:13] Step 1200: loss: 0.508326, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.006061
127
+ [2026-04-05 13:27:39] Step 1210: loss: 0.634969, grad_norm: 0.185547, learning_rate: 0.000050, epoch: 0.006111
128
+ [2026-04-05 13:29:05] Step 1220: loss: 0.454523, grad_norm: 0.191406, learning_rate: 0.000050, epoch: 0.006162
129
+ [2026-04-05 13:30:31] Step 1230: loss: 0.529709, grad_norm: 0.165039, learning_rate: 0.000050, epoch: 0.006212
130
+ [2026-04-05 13:31:57] Step 1240: loss: 0.654458, grad_norm: 0.217773, learning_rate: 0.000050, epoch: 0.006263
131
+ [2026-04-05 13:33:24] Step 1250: loss: 0.644524, grad_norm: 0.145508, learning_rate: 0.000050, epoch: 0.006313
132
+ [2026-04-05 13:34:50] Step 1260: loss: 0.558337, grad_norm: 0.186523, learning_rate: 0.000050, epoch: 0.006364
133
+ [2026-04-05 13:36:17] Step 1270: loss: 0.653074, grad_norm: 0.197266, learning_rate: 0.000050, epoch: 0.006414
134
+ [2026-04-05 13:37:44] Step 1280: loss: 0.553103, grad_norm: 0.476562, learning_rate: 0.000050, epoch: 0.006465
135
+ [2026-04-05 13:39:10] Step 1290: loss: 0.591599, grad_norm: 0.165039, learning_rate: 0.000050, epoch: 0.006515
136
+ [2026-04-05 13:40:37] Step 1300: loss: 0.455996, grad_norm: 0.163086, learning_rate: 0.000050, epoch: 0.006566
137
+ [2026-04-05 13:42:02] Step 1310: loss: 0.569381, grad_norm: 0.230469, learning_rate: 0.000050, epoch: 0.006616
138
+ [2026-04-05 13:43:28] Step 1320: loss: 0.583628, grad_norm: 0.169922, learning_rate: 0.000050, epoch: 0.006667
139
+ [2026-04-05 13:44:55] Step 1330: loss: 0.469083, grad_norm: 0.185547, learning_rate: 0.000050, epoch: 0.006717
140
+ [2026-04-05 13:46:22] Step 1340: loss: 0.524718, grad_norm: 0.180664, learning_rate: 0.000050, epoch: 0.006768
141
+ [2026-04-05 13:47:48] Step 1350: loss: 0.517907, grad_norm: 0.746094, learning_rate: 0.000050, epoch: 0.006818
142
+ [2026-04-05 13:49:14] Step 1360: loss: 0.545603, grad_norm: 0.306641, learning_rate: 0.000050, epoch: 0.006869
143
+ [2026-04-05 13:50:41] Step 1370: loss: 0.430360, grad_norm: 0.169922, learning_rate: 0.000050, epoch: 0.006919
144
+ [2026-04-05 13:52:06] Step 1380: loss: 0.564411, grad_norm: 0.535156, learning_rate: 0.000050, epoch: 0.006970
145
+ [2026-04-05 13:53:33] Step 1390: loss: 0.523255, grad_norm: 0.330078, learning_rate: 0.000050, epoch: 0.007020
146
+ [2026-04-05 13:55:00] Step 1400: loss: 0.616001, grad_norm: 0.470703, learning_rate: 0.000050, epoch: 0.007071
147
+ [2026-04-05 13:56:25] Step 1410: loss: 0.639578, grad_norm: 0.178711, learning_rate: 0.000050, epoch: 0.007121
148
+ [2026-04-05 13:57:50] Step 1420: loss: 0.471984, grad_norm: 0.310547, learning_rate: 0.000050, epoch: 0.007172
149
+ [2026-04-05 13:59:16] Step 1430: loss: 0.549651, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.007222
150
+ [2026-04-05 14:00:42] Step 1440: loss: 0.576872, grad_norm: 0.235352, learning_rate: 0.000050, epoch: 0.007273
151
+ [2026-04-05 14:02:08] Step 1450: loss: 0.552616, grad_norm: 0.242188, learning_rate: 0.000050, epoch: 0.007323
152
+ [2026-04-05 14:03:34] Step 1460: loss: 0.531245, grad_norm: 0.177734, learning_rate: 0.000050, epoch: 0.007374
153
+ [2026-04-05 14:05:01] Step 1470: loss: 1.899408, grad_norm: 0.574219, learning_rate: 0.000050, epoch: 0.007424
154
+ [2026-04-05 14:06:26] Step 1480: loss: 1.379459, grad_norm: 0.466797, learning_rate: 0.000050, epoch: 0.007475
155
+ [2026-04-05 14:07:52] Step 1490: loss: 1.148382, grad_norm: 0.281250, learning_rate: 0.000050, epoch: 0.007525
156
+ [2026-04-05 14:09:18] Step 1500: loss: 1.149607, grad_norm: 0.235352, learning_rate: 0.000050, epoch: 0.007576
157
+ [2026-04-05 14:10:45] Step 1510: loss: 1.059986, grad_norm: 0.345703, learning_rate: 0.000050, epoch: 0.007626
158
+ [2026-04-05 14:12:11] Step 1520: loss: 0.955449, grad_norm: 2.234375, learning_rate: 0.000050, epoch: 0.007677
159
+ [2026-04-05 14:13:37] Step 1530: loss: 1.007732, grad_norm: 0.380859, learning_rate: 0.000050, epoch: 0.007727
160
+ [2026-04-05 14:15:02] Step 1540: loss: 1.008875, grad_norm: 0.542969, learning_rate: 0.000050, epoch: 0.007778
161
+ [2026-04-05 14:16:28] Step 1550: loss: 1.136967, grad_norm: 6.312500, learning_rate: 0.000050, epoch: 0.007828
162
+ [2026-04-05 14:17:52] Step 1560: loss: 1.325755, grad_norm: 0.455078, learning_rate: 0.000050, epoch: 0.007879
163
+ [2026-04-05 14:19:18] Step 1570: loss: 1.099739, grad_norm: 0.410156, learning_rate: 0.000050, epoch: 0.007929
164
+ [2026-04-05 14:20:43] Step 1580: loss: 1.002961, grad_norm: 0.554688, learning_rate: 0.000050, epoch: 0.007980
165
+ [2026-04-05 14:22:09] Step 1590: loss: 1.057071, grad_norm: 0.250000, learning_rate: 0.000050, epoch: 0.008030
166
+ [2026-04-05 14:23:35] Step 1600: loss: 0.905516, grad_norm: 0.249023, learning_rate: 0.000050, epoch: 0.008081
167
+ [2026-04-05 14:25:01] Step 1610: loss: 1.050816, grad_norm: 0.337891, learning_rate: 0.000050, epoch: 0.008131
168
+ [2026-04-05 14:26:27] Step 1620: loss: 0.760539, grad_norm: 0.238281, learning_rate: 0.000050, epoch: 0.008182
169
+ [2026-04-05 14:27:54] Step 1630: loss: 0.866014, grad_norm: 0.337891, learning_rate: 0.000050, epoch: 0.008232
170
+ [2026-04-05 14:29:20] Step 1640: loss: 0.878096, grad_norm: 0.878906, learning_rate: 0.000050, epoch: 0.008283
171
+ [2026-04-05 14:30:46] Step 1650: loss: 0.899268, grad_norm: 0.198242, learning_rate: 0.000050, epoch: 0.008333
172
+ [2026-04-05 14:32:12] Step 1660: loss: 0.805451, grad_norm: 0.312500, learning_rate: 0.000050, epoch: 0.008384
173
+ [2026-04-05 14:33:37] Step 1670: loss: 0.875339, grad_norm: 0.226562, learning_rate: 0.000050, epoch: 0.008434
174
+ [2026-04-05 14:35:04] Step 1680: loss: 0.903578, grad_norm: 0.197266, learning_rate: 0.000050, epoch: 0.008485
175
+ [2026-04-05 14:36:30] Step 1690: loss: 0.893119, grad_norm: 0.500000, learning_rate: 0.000050, epoch: 0.008535
176
+ [2026-04-05 14:37:56] Step 1700: loss: 0.808819, grad_norm: 0.170898, learning_rate: 0.000050, epoch: 0.008586
177
+ [2026-04-05 14:39:21] Step 1710: loss: 0.837901, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.008636
178
+ [2026-04-05 14:40:48] Step 1720: loss: 0.789509, grad_norm: 0.226562, learning_rate: 0.000050, epoch: 0.008687
179
+ [2026-04-05 14:42:13] Step 1730: loss: 0.984356, grad_norm: 0.265625, learning_rate: 0.000050, epoch: 0.008737
180
+ [2026-04-05 14:43:38] Step 1740: loss: 1.009939, grad_norm: 0.218750, learning_rate: 0.000050, epoch: 0.008788
181
+ [2026-04-05 14:45:04] Step 1750: loss: 0.945137, grad_norm: 1.250000, learning_rate: 0.000050, epoch: 0.008838
182
+ [2026-04-05 14:46:30] Step 1760: loss: 0.892640, grad_norm: 0.200195, learning_rate: 0.000050, epoch: 0.008889
183
+ [2026-04-05 14:47:56] Step 1770: loss: 0.845068, grad_norm: 0.240234, learning_rate: 0.000050, epoch: 0.008939
184
+ [2026-04-05 14:49:20] Step 1780: loss: 1.132531, grad_norm: 0.652344, learning_rate: 0.000050, epoch: 0.008990
185
+ [2026-04-05 14:50:47] Step 1790: loss: 1.225483, grad_norm: 32.750000, learning_rate: 0.000050, epoch: 0.009040
186
+ [2026-04-05 14:52:13] Step 1800: loss: 1.753622, grad_norm: 0.304688, learning_rate: 0.000050, epoch: 0.009091
187
+ [2026-04-05 14:53:39] Step 1810: loss: 1.767740, grad_norm: 3.062500, learning_rate: 0.000050, epoch: 0.009141
188
+ [2026-04-05 14:55:05] Step 1820: loss: 1.830482, grad_norm: 2.359375, learning_rate: 0.000050, epoch: 0.009192
189
+ [2026-04-05 14:56:32] Step 1830: loss: 1.026241, grad_norm: 0.345703, learning_rate: 0.000050, epoch: 0.009242
190
+ [2026-04-05 14:57:58] Step 1840: loss: 1.049114, grad_norm: 0.237305, learning_rate: 0.000050, epoch: 0.009293
191
+ [2026-04-05 14:59:25] Step 1850: loss: 1.213499, grad_norm: 0.227539, learning_rate: 0.000050, epoch: 0.009343
192
+ [2026-04-05 15:00:51] Step 1860: loss: 1.006206, grad_norm: 0.355469, learning_rate: 0.000050, epoch: 0.009394
193
+ [2026-04-05 15:02:17] Step 1870: loss: 0.872990, grad_norm: 0.176758, learning_rate: 0.000050, epoch: 0.009444
194
+ [2026-04-05 15:03:42] Step 1880: loss: 0.959800, grad_norm: 0.357422, learning_rate: 0.000050, epoch: 0.009495
195
+ [2026-04-05 15:05:08] Step 1890: loss: 0.843189, grad_norm: 0.269531, learning_rate: 0.000050, epoch: 0.009545
196
+ [2026-04-05 15:06:33] Step 1900: loss: 0.732752, grad_norm: 0.384766, learning_rate: 0.000050, epoch: 0.009596
197
+ [2026-04-05 15:07:59] Step 1910: loss: 0.886395, grad_norm: 0.166016, learning_rate: 0.000050, epoch: 0.009646
198
+ [2026-04-05 15:09:25] Step 1920: loss: 0.814924, grad_norm: 0.179688, learning_rate: 0.000050, epoch: 0.009697
199
+ [2026-04-05 15:10:51] Step 1930: loss: 0.955270, grad_norm: 0.335938, learning_rate: 0.000050, epoch: 0.009747
200
+ [2026-04-05 15:12:18] Step 1940: loss: 1.084479, grad_norm: 0.250000, learning_rate: 0.000050, epoch: 0.009798
201
+ [2026-04-05 15:13:43] Step 1950: loss: 1.049752, grad_norm: 0.255859, learning_rate: 0.000050, epoch: 0.009848
202
+ [2026-04-05 15:15:10] Step 1960: loss: 0.852993, grad_norm: 0.197266, learning_rate: 0.000050, epoch: 0.009899
203
+ [2026-04-05 15:16:36] Step 1970: loss: 1.096658, grad_norm: 0.166016, learning_rate: 0.000050, epoch: 0.009949
204
+ [2026-04-05 15:18:03] Step 1980: loss: 0.874805, grad_norm: 0.217773, learning_rate: 0.000050, epoch: 0.010000
205
+ [2026-04-05 15:19:28] Step 1990: loss: 0.894336, grad_norm: 0.163086, learning_rate: 0.000050, epoch: 0.010051
206
+ [2026-04-05 15:20:54] Step 2000: loss: 0.869858, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.010101
207
+ [2026-04-05 15:22:20] Step 2010: loss: 0.753163, grad_norm: 0.175781, learning_rate: 0.000050, epoch: 0.010152
208
+ [2026-04-05 15:23:44] Step 2020: loss: 0.821059, grad_norm: 0.289062, learning_rate: 0.000050, epoch: 0.010202
209
+ [2026-04-05 15:25:09] Step 2030: loss: 0.929537, grad_norm: 0.326172, learning_rate: 0.000050, epoch: 0.010253
210
+ [2026-04-05 15:26:35] Step 2040: loss: 0.774003, grad_norm: 0.217773, learning_rate: 0.000050, epoch: 0.010303
211
+ [2026-04-05 15:28:00] Step 2050: loss: 0.576957, grad_norm: 0.189453, learning_rate: 0.000050, epoch: 0.010354
212
+ [2026-04-05 15:29:27] Step 2060: loss: 0.834949, grad_norm: 0.163086, learning_rate: 0.000050, epoch: 0.010404
213
+ [2026-04-05 15:30:52] Step 2070: loss: 0.714023, grad_norm: 0.213867, learning_rate: 0.000050, epoch: 0.010455
214
+ [2026-04-05 15:32:18] Step 2080: loss: 0.683500, grad_norm: 0.187500, learning_rate: 0.000050, epoch: 0.010505
215
+ [2026-04-05 15:33:44] Step 2090: loss: 0.759647, grad_norm: 0.220703, learning_rate: 0.000050, epoch: 0.010556
216
+ [2026-04-05 15:35:11] Step 2100: loss: 0.683842, grad_norm: 0.146484, learning_rate: 0.000050, epoch: 0.010606
217
+ [2026-04-05 15:36:37] Step 2110: loss: 0.823089, grad_norm: 0.138672, learning_rate: 0.000050, epoch: 0.010657
218
+ [2026-04-05 15:38:03] Step 2120: loss: 0.762830, grad_norm: 0.210938, learning_rate: 0.000050, epoch: 0.010707
219
+ [2026-04-05 15:39:28] Step 2130: loss: 0.739592, grad_norm: 0.182617, learning_rate: 0.000050, epoch: 0.010758
220
+ [2026-04-05 15:40:55] Step 2140: loss: 0.734296, grad_norm: 0.165039, learning_rate: 0.000050, epoch: 0.010808
221
+ [2026-04-05 15:42:22] Step 2150: loss: 0.685651, grad_norm: 0.217773, learning_rate: 0.000050, epoch: 0.010859
222
+ [2026-04-05 15:43:48] Step 2160: loss: 0.764905, grad_norm: 0.129883, learning_rate: 0.000050, epoch: 0.010909
223
+ [2026-04-05 15:45:14] Step 2170: loss: 0.686083, grad_norm: 0.140625, learning_rate: 0.000050, epoch: 0.010960
224
+ [2026-04-05 15:46:41] Step 2180: loss: 0.763166, grad_norm: 0.255859, learning_rate: 0.000050, epoch: 0.011010
225
+ [2026-04-05 15:48:06] Step 2190: loss: 0.669076, grad_norm: 0.127930, learning_rate: 0.000050, epoch: 0.011061
226
+ [2026-04-05 15:49:32] Step 2200: loss: 0.650153, grad_norm: 0.304688, learning_rate: 0.000050, epoch: 0.011111
227
+ [2026-04-05 15:50:58] Step 2210: loss: 0.832567, grad_norm: 0.210938, learning_rate: 0.000050, epoch: 0.011162
228
+ [2026-04-05 15:52:23] Step 2220: loss: 1.139757, grad_norm: 1.109375, learning_rate: 0.000050, epoch: 0.011212
229
+ [2026-04-05 15:53:47] Step 2230: loss: 1.250969, grad_norm: 0.484375, learning_rate: 0.000050, epoch: 0.011263
230
+ [2026-04-05 15:55:11] Step 2240: loss: 1.130311, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.011313
231
+ [2026-04-05 15:56:35] Step 2250: loss: 1.492599, grad_norm: 0.898438, learning_rate: 0.000050, epoch: 0.011364
232
+ [2026-04-05 15:57:59] Step 2260: loss: 1.048409, grad_norm: 0.773438, learning_rate: 0.000050, epoch: 0.011414
233
+ [2026-04-05 15:59:22] Step 2270: loss: 1.264527, grad_norm: 0.558594, learning_rate: 0.000050, epoch: 0.011465
234
+ [2026-04-05 16:00:46] Step 2280: loss: 1.209084, grad_norm: 0.226562, learning_rate: 0.000050, epoch: 0.011515
235
+ [2026-04-05 16:02:10] Step 2290: loss: 1.030489, grad_norm: 1.007812, learning_rate: 0.000050, epoch: 0.011566
236
+ [2026-04-05 16:03:33] Step 2300: loss: 1.380919, grad_norm: 1.054688, learning_rate: 0.000050, epoch: 0.011616
237
+ [2026-04-05 16:04:57] Step 2310: loss: 1.237099, grad_norm: 1.898438, learning_rate: 0.000050, epoch: 0.011667
238
+ [2026-04-05 16:06:21] Step 2320: loss: 1.099260, grad_norm: 1.437500, learning_rate: 0.000050, epoch: 0.011717
239
+ [2026-04-05 16:07:45] Step 2330: loss: 1.179968, grad_norm: 1.156250, learning_rate: 0.000050, epoch: 0.011768
240
+ [2026-04-05 16:09:09] Step 2340: loss: 1.348586, grad_norm: 0.777344, learning_rate: 0.000050, epoch: 0.011818
241
+ [2026-04-05 16:10:33] Step 2350: loss: 1.062583, grad_norm: 0.345703, learning_rate: 0.000050, epoch: 0.011869
242
+ [2026-04-05 16:11:56] Step 2360: loss: 0.927481, grad_norm: 0.582031, learning_rate: 0.000050, epoch: 0.011919
243
+ [2026-04-05 16:13:21] Step 2370: loss: 1.113601, grad_norm: 0.478516, learning_rate: 0.000050, epoch: 0.011970
244
+ [2026-04-05 16:14:45] Step 2380: loss: 0.967371, grad_norm: 0.464844, learning_rate: 0.000050, epoch: 0.012020
245
+ [2026-04-05 16:16:08] Step 2390: loss: 1.120241, grad_norm: 1.148438, learning_rate: 0.000050, epoch: 0.012071
246
+ [2026-04-05 16:17:32] Step 2400: loss: 0.954143, grad_norm: 0.542969, learning_rate: 0.000050, epoch: 0.012121
247
+ [2026-04-05 16:18:58] Step 2410: loss: 2.039123, grad_norm: 0.503906, learning_rate: 0.000050, epoch: 0.012172
248
+ [2026-04-05 16:20:23] Step 2420: loss: 1.816168, grad_norm: 0.742188, learning_rate: 0.000050, epoch: 0.012222
249
+ [2026-04-05 16:21:49] Step 2430: loss: 1.829216, grad_norm: 8.875000, learning_rate: 0.000050, epoch: 0.012273
250
+ [2026-04-05 16:23:13] Step 2440: loss: 2.097993, grad_norm: 0.468750, learning_rate: 0.000050, epoch: 0.012323
251
+ [2026-04-05 16:24:39] Step 2450: loss: 1.479318, grad_norm: 0.824219, learning_rate: 0.000050, epoch: 0.012374
252
+ [2026-04-05 16:26:06] Step 2460: loss: 1.397717, grad_norm: 0.285156, learning_rate: 0.000050, epoch: 0.012424
253
+ [2026-04-05 16:27:32] Step 2470: loss: 1.590504, grad_norm: 0.808594, learning_rate: 0.000050, epoch: 0.012475
254
+ [2026-04-05 16:28:57] Step 2480: loss: 1.398745, grad_norm: 0.480469, learning_rate: 0.000050, epoch: 0.012525
255
+ [2026-04-05 16:30:24] Step 2490: loss: 1.382736, grad_norm: 2.500000, learning_rate: 0.000050, epoch: 0.012576
256
+ [2026-04-05 16:31:50] Step 2500: loss: 1.278419, grad_norm: 0.371094, learning_rate: 0.000050, epoch: 0.012626
257
+ [2026-04-05 16:33:16] Step 2510: loss: 1.223915, grad_norm: 0.233398, learning_rate: 0.000050, epoch: 0.012677
258
+ [2026-04-05 16:34:42] Step 2520: loss: 1.601561, grad_norm: 1.093750, learning_rate: 0.000050, epoch: 0.012727
259
+ [2026-04-05 16:36:08] Step 2530: loss: 1.213593, grad_norm: 4.312500, learning_rate: 0.000050, epoch: 0.012778
260
+ [2026-04-05 16:37:35] Step 2540: loss: 1.188114, grad_norm: 1.085938, learning_rate: 0.000050, epoch: 0.012828
261
+ [2026-04-05 16:39:01] Step 2550: loss: 1.616557, grad_norm: 0.257812, learning_rate: 0.000050, epoch: 0.012879
262
+ [2026-04-05 16:40:27] Step 2560: loss: 1.487203, grad_norm: 0.380859, learning_rate: 0.000050, epoch: 0.012929
263
+ [2026-04-05 16:41:53] Step 2570: loss: 1.499329, grad_norm: 0.640625, learning_rate: 0.000050, epoch: 0.012980
264
+ [2026-04-05 16:43:19] Step 2580: loss: 1.215436, grad_norm: 0.392578, learning_rate: 0.000050, epoch: 0.013030
265
+ [2026-04-05 16:44:45] Step 2590: loss: 1.291605, grad_norm: 2.015625, learning_rate: 0.000050, epoch: 0.013081
266
+ [2026-04-05 16:46:11] Step 2600: loss: 1.156400, grad_norm: 0.550781, learning_rate: 0.000050, epoch: 0.013131
267
+ [2026-04-05 16:47:37] Step 2610: loss: 1.193227, grad_norm: 0.582031, learning_rate: 0.000050, epoch: 0.013182
268
+ [2026-04-05 16:49:04] Step 2620: loss: 1.147099, grad_norm: 0.204102, learning_rate: 0.000050, epoch: 0.013232
269
+ [2026-04-05 16:50:29] Step 2630: loss: 1.581532, grad_norm: 0.294922, learning_rate: 0.000050, epoch: 0.013283
270
+ [2026-04-05 16:51:55] Step 2640: loss: 1.038281, grad_norm: 0.229492, learning_rate: 0.000050, epoch: 0.013333
271
+ [2026-04-05 16:53:21] Step 2650: loss: 1.072556, grad_norm: 0.251953, learning_rate: 0.000050, epoch: 0.013384
272
+ [2026-04-05 16:54:47] Step 2660: loss: 1.094269, grad_norm: 0.196289, learning_rate: 0.000050, epoch: 0.013434
273
+ [2026-04-05 16:56:14] Step 2670: loss: 0.990274, grad_norm: 0.191406, learning_rate: 0.000050, epoch: 0.013485
274
+ [2026-04-05 16:57:40] Step 2680: loss: 1.059046, grad_norm: 0.302734, learning_rate: 0.000050, epoch: 0.013535
275
+ [2026-04-05 16:59:06] Step 2690: loss: 0.863274, grad_norm: 0.367188, learning_rate: 0.000050, epoch: 0.013586
276
+ [2026-04-05 17:00:32] Step 2700: loss: 0.899668, grad_norm: 0.159180, learning_rate: 0.000050, epoch: 0.013636
277
+ [2026-04-05 17:01:57] Step 2710: loss: 1.116811, grad_norm: 1.015625, learning_rate: 0.000050, epoch: 0.013687
278
+ [2026-04-05 17:03:23] Step 2720: loss: 1.039225, grad_norm: 0.200195, learning_rate: 0.000050, epoch: 0.013737
279
+ [2026-04-05 17:04:50] Step 2730: loss: 0.910703, grad_norm: 0.302734, learning_rate: 0.000050, epoch: 0.013788
280
+ [2026-04-05 17:06:15] Step 2740: loss: 0.952942, grad_norm: 0.257812, learning_rate: 0.000050, epoch: 0.013838
281
+ [2026-04-05 17:07:41] Step 2750: loss: 0.836150, grad_norm: 0.218750, learning_rate: 0.000050, epoch: 0.013889
282
+ [2026-04-05 17:09:05] Step 2760: loss: 1.159457, grad_norm: 0.414062, learning_rate: 0.000050, epoch: 0.013939
283
+ [2026-04-05 17:10:30] Step 2770: loss: 1.032300, grad_norm: 0.388672, learning_rate: 0.000050, epoch: 0.013990
284
+ [2026-04-05 17:11:53] Step 2780: loss: 1.089910, grad_norm: 0.427734, learning_rate: 0.000050, epoch: 0.014040
285
+ [2026-04-05 17:13:17] Step 2790: loss: 0.992994, grad_norm: 0.335938, learning_rate: 0.000050, epoch: 0.014091
286
+ [2026-04-05 17:14:41] Step 2800: loss: 0.949600, grad_norm: 0.310547, learning_rate: 0.000050, epoch: 0.014141
287
+ [2026-04-05 17:16:06] Step 2810: loss: 0.910935, grad_norm: 0.273438, learning_rate: 0.000050, epoch: 0.014192
288
+ [2026-04-05 17:17:30] Step 2820: loss: 0.931577, grad_norm: 0.246094, learning_rate: 0.000050, epoch: 0.014242
289
+ [2026-04-05 17:18:54] Step 2830: loss: 1.009028, grad_norm: 0.531250, learning_rate: 0.000050, epoch: 0.014293
290
+ [2026-04-05 17:20:18] Step 2840: loss: 1.071189, grad_norm: 0.337891, learning_rate: 0.000050, epoch: 0.014343
291
+ [2026-04-05 17:21:42] Step 2850: loss: 0.848302, grad_norm: 0.306641, learning_rate: 0.000050, epoch: 0.014394
292
+ [2026-04-05 17:23:06] Step 2860: loss: 0.861690, grad_norm: 0.300781, learning_rate: 0.000050, epoch: 0.014444
293
+ [2026-04-05 17:24:30] Step 2870: loss: 0.811077, grad_norm: 0.371094, learning_rate: 0.000050, epoch: 0.014495
294
+ [2026-04-05 17:25:55] Step 2880: loss: 0.871173, grad_norm: 0.239258, learning_rate: 0.000050, epoch: 0.014545
295
+ [2026-04-05 17:27:19] Step 2890: loss: 0.917228, grad_norm: 0.351562, learning_rate: 0.000050, epoch: 0.014596
296
+ [2026-04-05 17:28:43] Step 2900: loss: 0.786758, grad_norm: 0.261719, learning_rate: 0.000050, epoch: 0.014646
297
+ [2026-04-05 17:30:08] Step 2910: loss: 1.140635, grad_norm: 0.371094, learning_rate: 0.000050, epoch: 0.014697
298
+ [2026-04-05 17:31:33] Step 2920: loss: 1.568506, grad_norm: 1.101562, learning_rate: 0.000050, epoch: 0.014747
299
+ [2026-04-05 17:32:58] Step 2930: loss: 1.817063, grad_norm: 0.357422, learning_rate: 0.000050, epoch: 0.014798
300
+ [2026-04-05 17:34:23] Step 2940: loss: 1.432751, grad_norm: 0.707031, learning_rate: 0.000050, epoch: 0.014848
301
+ [2026-04-05 17:35:48] Step 2950: loss: 1.237652, grad_norm: 0.388672, learning_rate: 0.000050, epoch: 0.014899
302
+ [2026-04-05 17:37:13] Step 2960: loss: 1.362940, grad_norm: 0.503906, learning_rate: 0.000050, epoch: 0.014949
303
+ [2026-04-05 17:38:38] Step 2970: loss: 1.296589, grad_norm: 0.687500, learning_rate: 0.000050, epoch: 0.015000
304
+ [2026-04-05 17:40:03] Step 2980: loss: 1.269760, grad_norm: 0.281250, learning_rate: 0.000050, epoch: 0.015051
305
+ [2026-04-05 17:41:29] Step 2990: loss: 1.306033, grad_norm: 0.664062, learning_rate: 0.000050, epoch: 0.015101
306
+ [2026-04-05 17:42:53] Step 3000: loss: 1.112861, grad_norm: 0.361328, learning_rate: 0.000050, epoch: 0.015152
307
+ [2026-04-05 17:44:17] Step 3010: loss: 1.465074, grad_norm: 0.357422, learning_rate: 0.000050, epoch: 0.015202
308
+ [2026-04-05 17:45:43] Step 3020: loss: 1.325169, grad_norm: 0.945312, learning_rate: 0.000050, epoch: 0.015253
309
+ [2026-04-05 17:47:07] Step 3030: loss: 1.242030, grad_norm: 1.453125, learning_rate: 0.000050, epoch: 0.015303
310
+ [2026-04-05 17:48:31] Step 3040: loss: 1.179085, grad_norm: 0.640625, learning_rate: 0.000050, epoch: 0.015354
311
+ [2026-04-05 17:49:57] Step 3050: loss: 1.195576, grad_norm: 0.296875, learning_rate: 0.000050, epoch: 0.015404
312
+ [2026-04-05 17:51:22] Step 3060: loss: 1.088905, grad_norm: 0.423828, learning_rate: 0.000050, epoch: 0.015455
313
+ [2026-04-05 17:52:46] Step 3070: loss: 1.133894, grad_norm: 2.093750, learning_rate: 0.000050, epoch: 0.015505
314
+ [2026-04-05 17:54:12] Step 3080: loss: 1.182403, grad_norm: 0.226562, learning_rate: 0.000050, epoch: 0.015556
315
+ [2026-04-05 17:55:38] Step 3090: loss: 1.205568, grad_norm: 0.213867, learning_rate: 0.000050, epoch: 0.015606
316
+ [2026-04-05 17:57:04] Step 3100: loss: 0.972407, grad_norm: 0.398438, learning_rate: 0.000050, epoch: 0.015657
317
+ [2026-04-05 17:58:28] Step 3110: loss: 0.908546, grad_norm: 0.174805, learning_rate: 0.000050, epoch: 0.015707
318
+ [2026-04-05 17:59:53] Step 3120: loss: 1.111314, grad_norm: 0.231445, learning_rate: 0.000050, epoch: 0.015758
319
+ [2026-04-05 18:01:18] Step 3130: loss: 0.902070, grad_norm: 0.152344, learning_rate: 0.000050, epoch: 0.015808
320
+ [2026-04-05 18:02:42] Step 3140: loss: 0.900240, grad_norm: 0.263672, learning_rate: 0.000050, epoch: 0.015859
321
+ [2026-04-05 18:04:05] Step 3150: loss: 0.646932, grad_norm: 0.161133, learning_rate: 0.000050, epoch: 0.015909
322
+ [2026-04-05 18:05:30] Step 3160: loss: 1.058152, grad_norm: 0.247070, learning_rate: 0.000050, epoch: 0.015960
323
+ [2026-04-05 18:06:55] Step 3170: loss: 1.120416, grad_norm: 2.734375, learning_rate: 0.000050, epoch: 0.016010
324
+ [2026-04-05 18:08:20] Step 3180: loss: 1.347795, grad_norm: 0.410156, learning_rate: 0.000050, epoch: 0.016061
325
+ [2026-04-05 18:09:45] Step 3190: loss: 1.075504, grad_norm: 0.929688, learning_rate: 0.000050, epoch: 0.016111
326
+ [2026-04-05 18:11:09] Step 3200: loss: 0.600300, grad_norm: 0.151367, learning_rate: 0.000050, epoch: 0.016162
327
+ [2026-04-05 18:12:33] Step 3210: loss: 0.955885, grad_norm: 0.259766, learning_rate: 0.000050, epoch: 0.016212
328
+ [2026-04-05 18:13:57] Step 3220: loss: 0.891304, grad_norm: 2.765625, learning_rate: 0.000050, epoch: 0.016263
329
+ [2026-04-05 18:15:22] Step 3230: loss: 1.085288, grad_norm: 0.169922, learning_rate: 0.000050, epoch: 0.016313
330
+ [2026-04-05 18:16:46] Step 3240: loss: 0.953492, grad_norm: 0.300781, learning_rate: 0.000050, epoch: 0.016364
331
+ [2026-04-05 18:18:11] Step 3250: loss: 1.122248, grad_norm: 0.449219, learning_rate: 0.000050, epoch: 0.016414
332
+ [2026-04-05 18:19:36] Step 3260: loss: 0.880321, grad_norm: 0.291016, learning_rate: 0.000050, epoch: 0.016465
333
+ [2026-04-05 18:21:02] Step 3270: loss: 1.138729, grad_norm: 0.351562, learning_rate: 0.000050, epoch: 0.016515
334
+ [2026-04-05 18:22:27] Step 3280: loss: 0.897787, grad_norm: 0.224609, learning_rate: 0.000050, epoch: 0.016566
335
+ [2026-04-05 18:23:51] Step 3290: loss: 0.702576, grad_norm: 0.132812, learning_rate: 0.000050, epoch: 0.016616
336
+ [2026-04-05 18:25:15] Step 3300: loss: 0.795779, grad_norm: 0.193359, learning_rate: 0.000050, epoch: 0.016667
337
+ [2026-04-05 18:26:40] Step 3310: loss: 1.040128, grad_norm: 0.792969, learning_rate: 0.000050, epoch: 0.016717
338
+ [2026-04-05 18:28:05] Step 3320: loss: 1.010918, grad_norm: 0.209961, learning_rate: 0.000050, epoch: 0.016768
339
+ [2026-04-05 18:29:31] Step 3330: loss: 1.156274, grad_norm: 0.154297, learning_rate: 0.000050, epoch: 0.016818
340
+ [2026-04-05 18:30:56] Step 3340: loss: 0.944842, grad_norm: 0.449219, learning_rate: 0.000050, epoch: 0.016869
341
+ [2026-04-05 18:32:21] Step 3350: loss: 1.268639, grad_norm: 2.531250, learning_rate: 0.000050, epoch: 0.016919
342
+ [2026-04-05 18:33:45] Step 3360: loss: 0.813688, grad_norm: 3.671875, learning_rate: 0.000050, epoch: 0.016970
343
+ [2026-04-05 18:35:11] Step 3370: loss: 1.152398, grad_norm: 0.333984, learning_rate: 0.000050, epoch: 0.017020
344
+ [2026-04-05 18:36:36] Step 3380: loss: 0.974917, grad_norm: 0.769531, learning_rate: 0.000050, epoch: 0.017071
345
+ [2026-04-05 18:38:02] Step 3390: loss: 0.954514, grad_norm: 0.203125, learning_rate: 0.000050, epoch: 0.017121
346
+ [2026-04-05 18:39:27] Step 3400: loss: 1.091822, grad_norm: 0.197266, learning_rate: 0.000050, epoch: 0.017172
347
+ [2026-04-05 18:40:53] Step 3410: loss: 1.011484, grad_norm: 0.176758, learning_rate: 0.000050, epoch: 0.017222
348
+ [2026-04-05 18:42:17] Step 3420: loss: 0.701433, grad_norm: 0.328125, learning_rate: 0.000050, epoch: 0.017273
349
+ [2026-04-05 18:43:42] Step 3430: loss: 0.868405, grad_norm: 0.128906, learning_rate: 0.000050, epoch: 0.017323
350
+ [2026-04-05 18:45:08] Step 3440: loss: 1.028577, grad_norm: 0.231445, learning_rate: 0.000050, epoch: 0.017374
351
+ [2026-04-05 18:46:33] Step 3450: loss: 1.005750, grad_norm: 0.181641, learning_rate: 0.000050, epoch: 0.017424
352
+ [2026-04-05 18:47:59] Step 3460: loss: 1.001823, grad_norm: 0.294922, learning_rate: 0.000050, epoch: 0.017475
353
+ [2026-04-05 18:49:23] Step 3470: loss: 0.667828, grad_norm: 0.357422, learning_rate: 0.000050, epoch: 0.017525
354
+ [2026-04-05 18:50:49] Step 3480: loss: 0.835672, grad_norm: 1.437500, learning_rate: 0.000050, epoch: 0.017576
355
+ [2026-04-05 18:52:14] Step 3490: loss: 1.058652, grad_norm: 0.135742, learning_rate: 0.000050, epoch: 0.017626
356
+ [2026-04-05 18:53:40] Step 3500: loss: 1.062490, grad_norm: 0.194336, learning_rate: 0.000050, epoch: 0.017677
357
+ [2026-04-05 18:55:05] Step 3510: loss: 0.915247, grad_norm: 0.124023, learning_rate: 0.000050, epoch: 0.017727
358
+ [2026-04-05 18:56:29] Step 3520: loss: 0.586168, grad_norm: 0.073242, learning_rate: 0.000050, epoch: 0.017778
359
+ [2026-04-05 18:57:54] Step 3530: loss: 0.816989, grad_norm: 146.000000, learning_rate: 0.000050, epoch: 0.017828
360
+ [2026-04-05 18:59:19] Step 3540: loss: 0.540502, grad_norm: 0.119629, learning_rate: 0.000050, epoch: 0.017879
361
+ [2026-04-05 19:00:42] Step 3550: loss: 0.578981, grad_norm: 0.115234, learning_rate: 0.000050, epoch: 0.017929
362
+ [2026-04-05 19:02:08] Step 3560: loss: 0.675821, grad_norm: 0.154297, learning_rate: 0.000050, epoch: 0.017980
363
+ [2026-04-05 19:03:34] Step 3570: loss: 0.992571, grad_norm: 0.500000, learning_rate: 0.000050, epoch: 0.018030
364
+ [2026-04-05 19:04:59] Step 3580: loss: 0.860287, grad_norm: 0.225586, learning_rate: 0.000050, epoch: 0.018081
365
+ [2026-04-05 19:06:23] Step 3590: loss: 0.701336, grad_norm: 0.605469, learning_rate: 0.000050, epoch: 0.018131
366
+ [2026-04-05 19:07:49] Step 3600: loss: 0.786840, grad_norm: 0.232422, learning_rate: 0.000050, epoch: 0.018182
367
+ [2026-04-05 19:09:14] Step 3610: loss: 0.871714, grad_norm: 0.225586, learning_rate: 0.000050, epoch: 0.018232
368
+ [2026-04-05 19:10:39] Step 3620: loss: 0.835710, grad_norm: 0.217773, learning_rate: 0.000050, epoch: 0.018283
369
+ [2026-04-05 19:12:04] Step 3630: loss: 0.831312, grad_norm: 0.209961, learning_rate: 0.000050, epoch: 0.018333
370
+ [2026-04-05 19:13:29] Step 3640: loss: 1.358666, grad_norm: 0.867188, learning_rate: 0.000050, epoch: 0.018384
371
+ [2026-04-05 19:14:52] Step 3650: loss: 2.279568, grad_norm: 0.437500, learning_rate: 0.000050, epoch: 0.018434
372
+ [2026-04-05 19:16:16] Step 3660: loss: 0.685308, grad_norm: 0.328125, learning_rate: 0.000050, epoch: 0.018485
373
+ [2026-04-05 19:17:40] Step 3670: loss: 0.648882, grad_norm: 0.166016, learning_rate: 0.000050, epoch: 0.018535
374
+ [2026-04-05 19:19:06] Step 3680: loss: 0.948299, grad_norm: 0.435547, learning_rate: 0.000050, epoch: 0.018586
375
+ [2026-04-05 19:20:31] Step 3690: loss: 0.790501, grad_norm: 0.130859, learning_rate: 0.000050, epoch: 0.018636
376
+ [2026-04-05 19:21:55] Step 3700: loss: 0.744720, grad_norm: 0.230469, learning_rate: 0.000050, epoch: 0.018687
377
+ [2026-04-05 19:23:21] Step 3710: loss: 0.935014, grad_norm: 0.165039, learning_rate: 0.000050, epoch: 0.018737
378
+ [2026-04-05 19:24:46] Step 3720: loss: 0.820345, grad_norm: 0.302734, learning_rate: 0.000050, epoch: 0.018788
379
+ [2026-04-05 19:26:11] Step 3730: loss: 0.955610, grad_norm: 0.146484, learning_rate: 0.000050, epoch: 0.018838
380
+ [2026-04-05 19:27:34] Step 3740: loss: 0.571479, grad_norm: 0.375000, learning_rate: 0.000050, epoch: 0.018889
381
+ [2026-04-05 19:29:00] Step 3750: loss: 1.160526, grad_norm: 0.244141, learning_rate: 0.000050, epoch: 0.018939
382
+ [2026-04-05 19:30:24] Step 3760: loss: 0.704879, grad_norm: 0.075684, learning_rate: 0.000050, epoch: 0.018990
383
+ [2026-04-05 19:31:47] Step 3770: loss: 0.281452, grad_norm: 0.090332, learning_rate: 0.000050, epoch: 0.019040
384
+ [2026-04-05 19:33:12] Step 3780: loss: 0.948658, grad_norm: 1.109375, learning_rate: 0.000050, epoch: 0.019091
385
+ [2026-04-05 19:34:37] Step 3790: loss: 0.881313, grad_norm: 0.208984, learning_rate: 0.000050, epoch: 0.019141
386
+ [2026-04-05 19:36:03] Step 3800: loss: 0.818209, grad_norm: 0.257812, learning_rate: 0.000050, epoch: 0.019192
387
+ [2026-04-05 19:37:27] Step 3810: loss: 0.673398, grad_norm: 0.076172, learning_rate: 0.000050, epoch: 0.019242
388
+ [2026-04-05 19:38:50] Step 3820: loss: 0.505238, grad_norm: 0.582031, learning_rate: 0.000050, epoch: 0.019293
389
+ [2026-04-05 19:40:15] Step 3830: loss: 0.891517, grad_norm: 0.238281, learning_rate: 0.000050, epoch: 0.019343
390
+ [2026-04-05 19:41:40] Step 3840: loss: 0.937767, grad_norm: 0.176758, learning_rate: 0.000050, epoch: 0.019394
391
+ [2026-04-05 19:43:05] Step 3850: loss: 0.587611, grad_norm: 0.183594, learning_rate: 0.000050, epoch: 0.019444
392
+ [2026-04-05 19:44:30] Step 3860: loss: 1.020237, grad_norm: 0.235352, learning_rate: 0.000050, epoch: 0.019495
393
+ [2026-04-05 19:45:55] Step 3870: loss: 1.054631, grad_norm: 0.208984, learning_rate: 0.000050, epoch: 0.019545
394
+ [2026-04-05 19:47:21] Step 3880: loss: 0.901759, grad_norm: 0.139648, learning_rate: 0.000050, epoch: 0.019596
395
+ [2026-04-05 19:48:45] Step 3890: loss: 0.718227, grad_norm: 0.148438, learning_rate: 0.000050, epoch: 0.019646
396
+ [2026-04-05 19:50:09] Step 3900: loss: 0.997413, grad_norm: 1.695312, learning_rate: 0.000050, epoch: 0.019697
397
+ [2026-04-05 19:51:35] Step 3910: loss: 0.907274, grad_norm: 0.161133, learning_rate: 0.000050, epoch: 0.019747
398
+ [2026-04-05 19:53:00] Step 3920: loss: 0.738626, grad_norm: 0.090332, learning_rate: 0.000050, epoch: 0.019798
399
+ [2026-04-05 19:54:25] Step 3930: loss: 1.373290, grad_norm: 0.235352, learning_rate: 0.000050, epoch: 0.019848
400
+ [2026-04-05 19:55:50] Step 3940: loss: 0.769842, grad_norm: 0.210938, learning_rate: 0.000050, epoch: 0.019899
401
+ [2026-04-05 19:57:15] Step 3950: loss: 0.937164, grad_norm: 0.292969, learning_rate: 0.000050, epoch: 0.019949
402
+ [2026-04-05 19:58:41] Step 3960: loss: 0.940461, grad_norm: 0.330078, learning_rate: 0.000050, epoch: 0.020000
403
+ [2026-04-05 20:00:04] Step 3970: loss: 0.805276, grad_norm: 0.229492, learning_rate: 0.000050, epoch: 0.020051
404
+ [2026-04-05 20:01:28] Step 3980: loss: 0.602984, grad_norm: 0.306641, learning_rate: 0.000050, epoch: 0.020101
405
+ [2026-04-05 20:02:53] Step 3990: loss: 0.888927, grad_norm: 0.193359, learning_rate: 0.000050, epoch: 0.020152
406
+ [2026-04-05 20:04:19] Step 4000: loss: 1.048596, grad_norm: 0.523438, learning_rate: 0.000050, epoch: 0.020202
407
+ [2026-04-05 20:05:44] Step 4010: loss: 0.881884, grad_norm: 0.457031, learning_rate: 0.000050, epoch: 0.020253
408
+ [2026-04-05 20:07:09] Step 4020: loss: 0.909422, grad_norm: 0.183594, learning_rate: 0.000050, epoch: 0.020303
409
+ [2026-04-05 20:08:33] Step 4030: loss: 0.751946, grad_norm: 0.441406, learning_rate: 0.000050, epoch: 0.020354
410
+ [2026-04-05 20:09:59] Step 4040: loss: 0.816604, grad_norm: 0.182617, learning_rate: 0.000050, epoch: 0.020404
411
+ [2026-04-05 20:11:24] Step 4050: loss: 1.197823, grad_norm: 0.210938, learning_rate: 0.000050, epoch: 0.020455
412
+ [2026-04-05 20:12:50] Step 4060: loss: 1.082392, grad_norm: 0.828125, learning_rate: 0.000050, epoch: 0.020505
413
+ [2026-04-05 20:14:15] Step 4070: loss: 0.737737, grad_norm: 0.259766, learning_rate: 0.000050, epoch: 0.020556
414
+ [2026-04-05 20:15:40] Step 4080: loss: 0.567219, grad_norm: 0.316406, learning_rate: 0.000050, epoch: 0.020606
415
+ [2026-04-05 20:17:06] Step 4090: loss: 0.919196, grad_norm: 0.125977, learning_rate: 0.000050, epoch: 0.020657
416
+ [2026-04-05 20:18:31] Step 4100: loss: 1.070574, grad_norm: 0.175781, learning_rate: 0.000050, epoch: 0.020707
417
+ [2026-04-05 20:19:56] Step 4110: loss: 0.914139, grad_norm: 0.126953, learning_rate: 0.000050, epoch: 0.020758
418
+ [2026-04-05 20:21:20] Step 4120: loss: 0.658527, grad_norm: 0.242188, learning_rate: 0.000050, epoch: 0.020808
419
+ [2026-04-05 20:22:46] Step 4130: loss: 1.043363, grad_norm: 0.613281, learning_rate: 0.000050, epoch: 0.020859
420
+ [2026-04-05 20:24:09] Step 4140: loss: 0.686100, grad_norm: 0.133789, learning_rate: 0.000050, epoch: 0.020909
421
+ [2026-04-05 20:25:33] Step 4150: loss: 0.803859, grad_norm: 0.242188, learning_rate: 0.000050, epoch: 0.020960
422
+ [2026-04-05 20:26:58] Step 4160: loss: 0.858215, grad_norm: 0.102539, learning_rate: 0.000050, epoch: 0.021010
423
+ [2026-04-05 20:28:22] Step 4170: loss: 2.265225, grad_norm: 1.023438, learning_rate: 0.000050, epoch: 0.021061
424
+ [2026-04-05 20:29:46] Step 4180: loss: 1.579555, grad_norm: 0.605469, learning_rate: 0.000050, epoch: 0.021111
425
+ [2026-04-05 20:31:11] Step 4190: loss: 0.705098, grad_norm: 0.312500, learning_rate: 0.000050, epoch: 0.021162
426
+ [2026-04-05 20:32:36] Step 4200: loss: 0.678449, grad_norm: 0.257812, learning_rate: 0.000050, epoch: 0.021212
427
+ [2026-04-05 20:34:00] Step 4210: loss: 1.039099, grad_norm: 0.169922, learning_rate: 0.000050, epoch: 0.021263
428
+ [2026-04-05 20:35:25] Step 4220: loss: 0.897258, grad_norm: 0.392578, learning_rate: 0.000050, epoch: 0.021313
429
+ [2026-04-05 20:36:49] Step 4230: loss: 0.592581, grad_norm: 0.253906, learning_rate: 0.000050, epoch: 0.021364
430
+ [2026-04-05 20:38:14] Step 4240: loss: 0.581826, grad_norm: 0.414062, learning_rate: 0.000050, epoch: 0.021414
431
+ [2026-04-05 20:39:39] Step 4250: loss: 0.773074, grad_norm: 0.109863, learning_rate: 0.000050, epoch: 0.021465
432
+ [2026-04-05 20:41:04] Step 4260: loss: 0.772055, grad_norm: 0.092773, learning_rate: 0.000050, epoch: 0.021515
433
+ [2026-04-05 20:42:29] Step 4270: loss: 0.519814, grad_norm: 0.173828, learning_rate: 0.000050, epoch: 0.021566
434
+ [2026-04-05 20:43:54] Step 4280: loss: 0.743214, grad_norm: 0.206055, learning_rate: 0.000050, epoch: 0.021616
435
+ [2026-04-05 20:45:19] Step 4290: loss: 0.712543, grad_norm: 0.192383, learning_rate: 0.000050, epoch: 0.021667
436
+ [2026-04-05 20:46:45] Step 4300: loss: 0.858545, grad_norm: 0.112793, learning_rate: 0.000050, epoch: 0.021717
437
+ [2026-04-05 20:48:10] Step 4310: loss: 0.812474, grad_norm: 0.244141, learning_rate: 0.000050, epoch: 0.021768
438
+ [2026-04-05 20:49:34] Step 4320: loss: 0.804857, grad_norm: 0.174805, learning_rate: 0.000050, epoch: 0.021818
439
+ [2026-04-05 20:50:59] Step 4330: loss: 0.676323, grad_norm: 0.060547, learning_rate: 0.000050, epoch: 0.021869
440
+ [2026-04-05 20:52:25] Step 4340: loss: 0.810262, grad_norm: 0.476562, learning_rate: 0.000050, epoch: 0.021919
441
+ [2026-04-05 20:53:49] Step 4350: loss: 0.678963, grad_norm: 1.210938, learning_rate: 0.000050, epoch: 0.021970
442
+ [2026-04-05 20:55:15] Step 4360: loss: 0.843389, grad_norm: 0.096680, learning_rate: 0.000050, epoch: 0.022020
443
+ [2026-04-05 20:56:40] Step 4370: loss: 0.557602, grad_norm: 0.070312, learning_rate: 0.000050, epoch: 0.022071
444
+ [2026-04-05 20:58:04] Step 4380: loss: 0.352494, grad_norm: 0.298828, learning_rate: 0.000050, epoch: 0.022121
445
+ [2026-04-05 20:59:28] Step 4390: loss: 0.624163, grad_norm: 0.472656, learning_rate: 0.000050, epoch: 0.022172
446
+ [2026-04-05 21:00:53] Step 4400: loss: 0.594077, grad_norm: 0.093262, learning_rate: 0.000050, epoch: 0.022222
447
+ [2026-04-05 21:02:18] Step 4410: loss: 0.878214, grad_norm: 0.330078, learning_rate: 0.000050, epoch: 0.022273
448
+ [2026-04-05 21:03:44] Step 4420: loss: 0.740005, grad_norm: 0.503906, learning_rate: 0.000050, epoch: 0.022323
449
+ [2026-04-05 21:05:08] Step 4430: loss: 0.971799, grad_norm: 0.562500, learning_rate: 0.000050, epoch: 0.022374
450
+ [2026-04-05 21:06:33] Step 4440: loss: 0.924411, grad_norm: 0.261719, learning_rate: 0.000050, epoch: 0.022424
451
+ [2026-04-05 21:07:59] Step 4450: loss: 1.042776, grad_norm: 0.140625, learning_rate: 0.000050, epoch: 0.022475
452
+ [2026-04-05 21:09:24] Step 4460: loss: 0.742720, grad_norm: 0.126953, learning_rate: 0.000050, epoch: 0.022525
453
+ [2026-04-05 21:10:49] Step 4470: loss: 0.903004, grad_norm: 0.233398, learning_rate: 0.000050, epoch: 0.022576
454
+ [2026-04-05 21:12:13] Step 4480: loss: 0.624388, grad_norm: 0.161133, learning_rate: 0.000050, epoch: 0.022626
455
+ [2026-04-05 21:13:39] Step 4490: loss: 0.894780, grad_norm: 0.178711, learning_rate: 0.000050, epoch: 0.022677
456
+ [2026-04-05 21:15:04] Step 4500: loss: 0.783386, grad_norm: 0.275391, learning_rate: 0.000050, epoch: 0.022727
457
+ [2026-04-05 21:16:30] Step 4510: loss: 0.852587, grad_norm: 0.182617, learning_rate: 0.000050, epoch: 0.022778
458
+ [2026-04-05 21:17:56] Step 4520: loss: 0.751036, grad_norm: 0.124023, learning_rate: 0.000050, epoch: 0.022828
459
+ [2026-04-05 21:19:21] Step 4530: loss: 0.812920, grad_norm: 0.808594, learning_rate: 0.000050, epoch: 0.022879
460
+ [2026-04-05 21:20:47] Step 4540: loss: 0.928886, grad_norm: 0.201172, learning_rate: 0.000050, epoch: 0.022929
461
+ [2026-04-05 21:22:12] Step 4550: loss: 0.699640, grad_norm: 0.117188, learning_rate: 0.000050, epoch: 0.022980
462
+ [2026-04-05 21:23:35] Step 4560: loss: 0.396365, grad_norm: 0.096191, learning_rate: 0.000050, epoch: 0.023030
463
+ [2026-04-05 21:24:59] Step 4570: loss: 0.460258, grad_norm: 0.089844, learning_rate: 0.000050, epoch: 0.023081
464
+ [2026-04-05 21:26:24] Step 4580: loss: 0.803078, grad_norm: 0.455078, learning_rate: 0.000050, epoch: 0.023131
465
+ [2026-04-05 21:27:48] Step 4590: loss: 0.623093, grad_norm: 0.175781, learning_rate: 0.000050, epoch: 0.023182
466
+ [2026-04-05 21:29:14] Step 4600: loss: 1.008995, grad_norm: 0.166016, learning_rate: 0.000050, epoch: 0.023232
467
+ [2026-04-05 21:30:39] Step 4610: loss: 1.163760, grad_norm: 0.181641, learning_rate: 0.000050, epoch: 0.023283
468
+ [2026-04-05 21:32:03] Step 4620: loss: 0.569889, grad_norm: 0.135742, learning_rate: 0.000050, epoch: 0.023333
469
+ [2026-04-05 21:33:29] Step 4630: loss: 1.077146, grad_norm: 0.158203, learning_rate: 0.000050, epoch: 0.023384
470
+ [2026-04-05 21:34:54] Step 4640: loss: 0.789613, grad_norm: 0.316406, learning_rate: 0.000050, epoch: 0.023434
471
+ [2026-04-05 21:36:19] Step 4650: loss: 0.768905, grad_norm: 0.185547, learning_rate: 0.000050, epoch: 0.023485
472
+ [2026-04-05 21:37:42] Step 4660: loss: 0.265431, grad_norm: 0.196289, learning_rate: 0.000050, epoch: 0.023535
473
+ [2026-04-05 21:39:06] Step 4670: loss: 0.509521, grad_norm: 0.174805, learning_rate: 0.000050, epoch: 0.023586
474
+ [2026-04-05 21:40:32] Step 4680: loss: 0.919802, grad_norm: 0.275391, learning_rate: 0.000050, epoch: 0.023636
475
+ [2026-04-05 21:41:57] Step 4690: loss: 0.862871, grad_norm: 0.269531, learning_rate: 0.000050, epoch: 0.023687
476
+ [2026-04-05 21:43:21] Step 4700: loss: 0.451978, grad_norm: 0.371094, learning_rate: 0.000050, epoch: 0.023737
477
+ [2026-04-05 21:44:45] Step 4710: loss: 0.829016, grad_norm: 0.146484, learning_rate: 0.000050, epoch: 0.023788
478
+ [2026-04-05 21:46:11] Step 4720: loss: 0.757640, grad_norm: 0.145508, learning_rate: 0.000050, epoch: 0.023838
479
+ [2026-04-05 21:47:36] Step 4730: loss: 0.821686, grad_norm: 0.173828, learning_rate: 0.000050, epoch: 0.023889
480
+ [2026-04-05 21:49:01] Step 4740: loss: 0.869125, grad_norm: 0.494141, learning_rate: 0.000050, epoch: 0.023939
481
+ [2026-04-05 21:50:26] Step 4750: loss: 0.680020, grad_norm: 0.092285, learning_rate: 0.000050, epoch: 0.023990
482
+ [2026-04-05 21:51:51] Step 4760: loss: 0.594586, grad_norm: 0.221680, learning_rate: 0.000050, epoch: 0.024040
483
+ [2026-04-05 21:53:16] Step 4770: loss: 0.768582, grad_norm: 0.179688, learning_rate: 0.000050, epoch: 0.024091
484
+ [2026-04-05 21:54:40] Step 4780: loss: 0.582672, grad_norm: 0.112793, learning_rate: 0.000050, epoch: 0.024141
485
+ [2026-04-05 21:56:06] Step 4790: loss: 0.791890, grad_norm: 0.120117, learning_rate: 0.000050, epoch: 0.024192
486
+ [2026-04-05 21:57:31] Step 4800: loss: 0.806308, grad_norm: 0.111328, learning_rate: 0.000050, epoch: 0.024242
487
+ [2026-04-05 21:58:56] Step 4810: loss: 0.739746, grad_norm: 0.148438, learning_rate: 0.000050, epoch: 0.024293
488
+ [2026-04-05 22:00:22] Step 4820: loss: 0.762415, grad_norm: 0.106445, learning_rate: 0.000050, epoch: 0.024343
489
+ [2026-04-05 22:01:47] Step 4830: loss: 0.877260, grad_norm: 0.212891, learning_rate: 0.000050, epoch: 0.024394
490
+ [2026-04-05 22:03:12] Step 4840: loss: 0.674090, grad_norm: 0.147461, learning_rate: 0.000050, epoch: 0.024444
491
+ [2026-04-05 22:04:36] Step 4850: loss: 0.761976, grad_norm: 0.124512, learning_rate: 0.000050, epoch: 0.024495
492
+ [2026-04-05 22:06:00] Step 4860: loss: 0.420770, grad_norm: 0.209961, learning_rate: 0.000050, epoch: 0.024545
493
+ [2026-04-05 22:07:25] Step 4870: loss: 0.853558, grad_norm: 0.146484, learning_rate: 0.000050, epoch: 0.024596
494
+ [2026-04-05 22:08:49] Step 4880: loss: 0.519837, grad_norm: 0.161133, learning_rate: 0.000050, epoch: 0.024646
495
+ [2026-04-05 22:10:14] Step 4890: loss: 0.812432, grad_norm: 0.139648, learning_rate: 0.000050, epoch: 0.024697
496
+ [2026-04-05 22:11:39] Step 4900: loss: 0.782813, grad_norm: 0.104004, learning_rate: 0.000050, epoch: 0.024747
497
+ [2026-04-05 22:13:05] Step 4910: loss: 0.806325, grad_norm: 0.156250, learning_rate: 0.000050, epoch: 0.024798
498
+ [2026-04-05 22:14:30] Step 4920: loss: 0.727326, grad_norm: 0.122559, learning_rate: 0.000050, epoch: 0.024848
499
+ [2026-04-05 22:15:54] Step 4930: loss: 0.891241, grad_norm: 0.261719, learning_rate: 0.000050, epoch: 0.024899
500
+ [2026-04-05 22:17:18] Step 4940: loss: 0.880500, grad_norm: 0.139648, learning_rate: 0.000050, epoch: 0.024949
501
+ [2026-04-05 22:18:43] Step 4950: loss: 0.869738, grad_norm: 0.621094, learning_rate: 0.000050, epoch: 0.025000
502
+ [2026-04-05 22:20:07] Step 4960: loss: 0.483802, grad_norm: 0.279297, learning_rate: 0.000050, epoch: 0.025051
503
+ [2026-04-05 22:21:32] Step 4970: loss: 0.424717, grad_norm: 0.068359, learning_rate: 0.000050, epoch: 0.025101
504
+ [2026-04-05 22:22:57] Step 4980: loss: 0.770722, grad_norm: 0.153320, learning_rate: 0.000050, epoch: 0.025152
505
+ [2026-04-05 22:24:21] Step 4990: loss: 0.476020, grad_norm: 0.094727, learning_rate: 0.000050, epoch: 0.025202
506
+ [2026-04-05 22:25:46] Step 5000: loss: 0.657169, grad_norm: 0.111816, learning_rate: 0.000050, epoch: 0.025253
507
+ [2026-04-05 22:28:18] Step 5010: loss: 0.819577, grad_norm: 0.121094, learning_rate: 0.000050, epoch: 0.025303