Update README.md

README.md

@@ -26,7 +26,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # vqa
 
-This model is a fine-tuned version of [unc-nlp/lxmert-base-uncased](https://huggingface.co/unc-nlp/lxmert-base-uncased) on the Graphcore/vqa-lxmert dataset.
+This model is a fine-tuned version of [unc-nlp/lxmert-base-uncased](https://huggingface.co/unc-nlp/lxmert-base-uncased) on the [Graphcore/vqa-lxmert](https://huggingface.co/datasets/Graphcore/vqa-lxmert) dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.0009
 - Accuracy: 0.7242
@@ -41,10 +41,39 @@ More information needed
 
 ## Training and evaluation data
 
-More information needed
+[Graphcore/vqa-lxmert](https://huggingface.co/datasets/Graphcore/vqa-lxmert) dataset
 
 ## Training procedure
 
+Trained on 16 Graphcore Mk2 IPUs using [optimum-graphcore](https://github.com/huggingface/optimum-graphcore).
+
+Command line:
+
+```
+python examples/language-modeling/run_clm.py \
+  --model_name_or_path gpt2 \
+  --ipu_config_name Graphcore/gpt2-small-ipu \
+  --dataset_name wikitext \
+  --dataset_config_name wikitext-103-raw-v1 \
+  --do_train \
+  --do_eval \
+  --num_train_epochs 10 \
+  --dataloader_num_workers 64 \
+  --per_device_train_batch_size 1 \
+  --per_device_eval_batch_size 1 \
+  --gradient_accumulation_steps 128 \
+  --output_dir /tmp/clm_output \
+  --logging_steps 5 \
+  --learning_rate 1e-5 \
+  --lr_scheduler_type linear \
+  --loss_scaling 16384 \
+  --weight_decay 0.01 \
+  --warmup_ratio 0.1 \
+  --ipu_config_overrides="embedding_serialization_factor=4,optimizer_state_offchip=true,inference_device_iterations=5" \
+  --dataloader_drop_last \
+  --pod_type pod16
+```
+
 ### Training hyperparameters
 
 The following hyperparameters were used during training: