Training in progress, step 500

config.json

@@ -5,7 +5,7 @@
   "add_bias_logits": false,
   "add_final_layer_norm": false,
   "architectures": [
-    "PipelinedBartForConditionalGeneration"
+    "PoptorchPipelinedBartForConditionalGeneration"
   ],
   "attention_dropout": 0.1,
   "bos_token_id": 0,

ipu_config.json

@@ -7,7 +7,7 @@
   "execute_encoder_on_cpu_for_generation": false,
   "gradient_accumulation_steps": 128,
   "inference_device_iterations": 6,
-  "inference_replication_factor": 4,
+  "inference_replication_factor": 1,
   "ipus_per_replica": 4,
   "layers_per_ipu": [
     0,
@@ -25,8 +25,8 @@
   "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 sha256:5187b7724855d803586b9715d32e4c6c8f7579377b719b48fb6d137aad166753
+oid sha256:884fbc940cd53ff641c808384929b4e2902ce4541b38457b090b52fe2874f0af
 size 279024057

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3123096a95a2e86c9ed51435f10bf85c58130ef41ec57a9517702e555c9c5b46
+oid sha256:4134102b42655682ff52c24e250b07795be661936bf3e898f217579d9a64a3f1
 size 2863