Upload log_ts_f_rep.txt with huggingface_hub
Browse files- log_ts_f_rep.txt +15 -0
log_ts_f_rep.txt
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
/tmp/tiny_train.py:237: UserWarning: Converting a tensor with requires_grad=True to a scalar may lead to unexpected behavior.
|
| 2 |
+
Consider using tensor.detach() first. (Triggered internally at /pytorch/torch/csrc/autograd/generated/python_variable_methods.cpp:836.)
|
| 3 |
+
total += float(l) * w
|
| 4 |
+
batch ready
|
| 5 |
+
10:25:25 epoch 0 start
|
| 6 |
+
10:25:43 step 78100 loss=2.5604 tok=1279.6M dt=18
|
| 7 |
+
10:25:51 step 78200 loss=2.4973 tok=1281.2M dt=26
|
| 8 |
+
10:25:58 step 78300 loss=2.5199 tok=1282.9M dt=33
|
| 9 |
+
10:26:06 step 78400 loss=2.4135 tok=1284.5M dt=41
|
| 10 |
+
10:26:14 step 78500 loss=2.4757 tok=1286.1M dt=48
|
| 11 |
+
10:26:21 step 78600 loss=2.3928 tok=1287.8M dt=56
|
| 12 |
+
10:26:29 step 78700 loss=2.4895 tok=1289.4M dt=63
|
| 13 |
+
10:26:37 step 78800 loss=2.3867 tok=1291.1M dt=71
|
| 14 |
+
10:26:45 step 78900 loss=2.3418 tok=1292.7M dt=79
|
| 15 |
+
10:26:53 step 79000 loss=2.4164 tok=1294.3M dt=87
|