Training in progress, epoch 1

ipu_config.json

@@ -7,7 +7,7 @@
  "execute_encoder_on_cpu_for_generation": false,
  "gradient_accumulation_steps": 32,
  "inference_device_iterations": 4,
- "inference_replication_factor": 4,
+ "inference_replication_factor": 1,
  "ipus_per_replica": 4,
  "layers_per_ipu": [
  3,
@@ -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:f8dd25d48d050729560aff2edd3a8804772060d7e9c67a560833f3cb4ba5d53e
+oid sha256:5aa5114d9beb6981467aaefcadf8c245dd4bb6510d47128c51f5b5588da2d908
 size 171676081

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:49bd008a209ddce49e35c3dd76836710e5879e82fe345784e53aca4f80a0f469
+oid sha256:f18137acf6e76f22de30f385045a33c51c7e4a1bccce77c50db07e1baaab2313
 size 2799