lysandre HF staff commited on
Commit
b279e8a
1 Parent(s): b8df249

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -1
README.md CHANGED
@@ -77,7 +77,10 @@ The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total)
77
  of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer
78
  used is Adam with a learning rate of 1e-4, \\(\beta_{1} = 0.9\\) and \\(\beta_{2} = 0.999\\), a weight decay of 0.01,
79
  learning rate warmup for 10,000 steps and linear decay of the learning rate after.
80
- ### Fine-tuningAfter pre-training, this model was fine-tuned on the SQuAD dataset with one of our fine-tuning scripts. In order to reproduce the training, you may use the following command:
 
 
 
81
  ```
82
  python -m torch.distributed.launch --nproc_per_node=8 ./examples/question-answering/run_qa.py \
83
  --model_name_or_path bert-large-uncased-whole-word-masking \
 
77
  of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer
78
  used is Adam with a learning rate of 1e-4, \\(\beta_{1} = 0.9\\) and \\(\beta_{2} = 0.999\\), a weight decay of 0.01,
79
  learning rate warmup for 10,000 steps and linear decay of the learning rate after.
80
+
81
+ ### Fine-tuning
82
+
83
+ After pre-training, this model was fine-tuned on the SQuAD dataset with one of our fine-tuning scripts. In order to reproduce the training, you may use the following command:
84
  ```
85
  python -m torch.distributed.launch --nproc_per_node=8 ./examples/question-answering/run_qa.py \
86
  --model_name_or_path bert-large-uncased-whole-word-masking \