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:c2cf307fc5ea138e2e0617eeed4767ff50620c2f721312075bdb74578d33ff28
+oid sha256:7538f5a04506a8684fd27983c2c8d595dd83f1a1f7587cb56bf13ef83959c451
 size 171676081

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e628e21e4268ccfe90a6dae29c0d8344916352a64f5a8655776f4097c76a4605
+oid sha256:81cc8d36f844c60508fa8385789833b2152fabcbf79b7951a6d46f6a9dcb645d
 size 2799