Training in progress, step 500

config.json

@@ -33,7 +33,7 @@
     }
   },
   "torch_dtype": "float16",
-  "transformers_version": "4.18.0",
+  "transformers_version": "4.20.1",
   "use_cache": true,
   "vocab_size": 50257
 }

ipu_config.json

@@ -1,10 +1,9 @@
 {
-  "decompose_grad_sum": false,
+  "auto_loss_scaling": false,
   "device_iterations": 2,
   "embedding_serialization_factor": 4,
-  "enable_half_first_order_momentum": true,
   "enable_half_partials": true,
-  "executable_cache_dir": "./exe_cache",
+  "executable_cache_dir": "/tmp/exe_cache",
   "execute_encoder_on_cpu_for_generation": false,
   "gradient_accumulation_steps": 16,
   "inference_device_iterations": 5,
@@ -18,14 +17,12 @@
   ],
   "matmul_proportion": 0.25,
   "optimizer_state_offchip": true,
-  "optimum_version": "1.1.1",
+  "optimum_version": "1.5.1",
   "output_mode": "final",
-  "profile_dir": "",
   "recompute_checkpoint_every_layer": true,
   "replicated_tensor_sharding": false,
   "replication_factor": 1,
   "seed": 42,
   "sharded_execution_for_inference": false,
-  "transformers_version": "4.18.0",
-  "use_popdist": false
+  "transformers_version": "4.20.1"
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2ec41fd9943e837076f0f51e71ade17eb70d2dddb442c721514ecb861ec5d721
+oid sha256:98be7a37645b50ad99a01b1d44d9ab4abd27738bff597b76c13b02edde1f3ed6
 size 261514473

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:388d799c1733715ab6da9bab9bd736a7df5fcaaafd9649521d4019e8b3b08e3c
-size 2671
+oid sha256:48abb1513a620a1159d070ffd661e2132964549a3d5f8120b3953f4519b604ca
+size 2735