Training in progress, epoch 1

ipu_config.json

@@ -7,7 +7,7 @@
   "execute_encoder_on_cpu_for_generation": false,
   "gradient_accumulation_steps": 16,
   "inference_device_iterations": 5,
-  "inference_replication_factor": 4,
+  "inference_replication_factor": 1,
   "ipus_per_replica": 4,
   "layers_per_ipu": [
     0,
@@ -22,8 +22,8 @@
   "optimum_version": "1.5.1",
   "output_mode": "final",
   "recompute_checkpoint_every_layer": true,
-  "replicated_tensor_sharding": true,
-  "replication_factor": 4,
+  "replicated_tensor_sharding": false,
+  "replication_factor": 1,
   "seed": 42,
   "sharded_execution_for_inference": false,
   "transformers_version": "4.20.1"

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b97efcdc791c28fddd5c32353aa6b28a1afd148815a025316b58fe1ff78ec85
+oid sha256:83b95d81f991aae1a0a57465f9338958f3e4dced35b6aa7db4bbd9ffe10a6bcc
 size 249363181

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b9cf320258f1610a60d9047503cd02ccbb53e03d9e4f8c0392a96167c4d708d2
+oid sha256:a519c77366311499578be1e726e76df5876b5d2942e14b7d2eda0bcc14039021
 size 2799