Training in progress, step 20000

config.json

@@ -6,10 +6,10 @@
   "d_kv": 64,
   "d_model": 768,
   "decoder_input_size": 1024,
-  "decoder_start_token_id": 0,
+  "decoder_start_token_id": 1025,
   "dense_act_fn": "relu",
   "dropout_rate": 0.2,
-  "eos_token_id": 1,
+  "eos_token_id": 1026,
   "feed_forward_proj": "relu",
   "initializer_factor": 1.0,
   "is_encoder_decoder": true,
@@ -22,11 +22,11 @@
   "num_decoder_layers": 10,
   "num_heads": 12,
   "num_layers": 10,
-  "pad_token_id": 0,
+  "pad_token_id": 1025,
   "relative_attention_max_distance": 128,
   "relative_attention_num_buckets": 32,
   "torch_dtype": "float32",
   "transformers_version": "4.26.0",
   "use_cache": true,
-  "vocab_size": 1024
+  "vocab_size": 1026
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:89f2ceb86af010d7ff89d4e2fd3a380a262921099e121e953ace9dd84f8994ed
-size 475241897
+oid sha256:5e949e5a9b8a4459672d5de8e89eeeb8edae69ab5b20d8cdffe462a3f62790e9
+size 475248041

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5d5c27846b17bf0edd60977e0796de118903abaadf8d79f83ad546f2f0500235
+oid sha256:f32179c0b62c57cde30389ba5eb01df73172b6f2b9c89b7a253c0cd58a201c4c
 size 3451