Training in progress, step 500

.gitignore

@@ -0,0 +1 @@
+checkpoint-*/

config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "roberta-base",
+  "architectures": [
+    "PoptorchPipelinedRobertaForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float16",
+  "transformers_version": "4.20.1",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

ipu_config.json

@@ -0,0 +1,30 @@
+{
+  "auto_loss_scaling": false,
+  "device_iterations": 1,
+  "embedding_serialization_factor": 0,
+  "enable_half_partials": true,
+  "executable_cache_dir": "/tmp/lmcache/",
+  "execute_encoder_on_cpu_for_generation": false,
+  "gradient_accumulation_steps": 16,
+  "inference_device_iterations": 5,
+  "inference_replication_factor": 4,
+  "ipus_per_replica": 4,
+  "layers_per_ipu": [
+    0,
+    4,
+    4,
+    4
+  ],
+  "matmul_proportion": [
+    0.25
+  ],
+  "optimizer_state_offchip": true,
+  "optimum_version": "1.5.1",
+  "output_mode": "final",
+  "recompute_checkpoint_every_layer": true,
+  "replicated_tensor_sharding": true,
+  "replication_factor": 4,
+  "seed": 42,
+  "sharded_execution_for_inference": false,
+  "transformers_version": "4.20.1"
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c46b9dce35d1478d85575c0ca4ba90750a525885439356a8e3d59391161d0b3f
+size 249463979

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6aeab10e041217dddff8fc1e0aed251579109410d0e5cd59d0cf23b3e835f7ef
+size 2735