Training in progress, step 500

config.json

@@ -1,7 +1,7 @@
 {
   "_name_or_path": "t5-small",
   "architectures": [
-    "PipelinedT5ForConditionalGeneration"
+    "PoptorchPipelinedT5ForConditionalGeneration"
   ],
   "d_ff": 2048,
   "d_kv": 64,

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,
@@ -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 sha256:40c97cabd74f9b644ac40d65c3f3ab45f006ef8c78771900288aee31ef9a6f1a
+oid sha256:8e8eac098140367a0fba1d56738c96b0c4c1a7187074a8a6d7d5bfaed35afd88
 size 121055163

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fdc004737cc12876f67d3c4fcd973cd086389ae76352752738debbe5baebd259
+oid sha256:b39b8d2451a66013ad718ab0cf6f9124df1f0733b578b11613d6fbb094ec3154
 size 2863