Upload T5ForConditionalGeneration

config.json

@@ -9,6 +9,7 @@
   "d_model": 512,
   "debug_mode": false,
   "decoder_start_token_id": 0,
+  "dense_act_fn": "relu",
   "description": "optimal joint config from sweep, mainly for qanom",
   "dir_switch": "joint_optimal",
   "do_eval_on": "validation",
@@ -21,6 +22,7 @@
   "gradient_accumulation_steps": 14,
   "initializer_factor": 1.0,
   "is_encoder_decoder": true,
+  "is_gated_act": false,
   "layer_norm_epsilon": 1e-06,
   "learning_rate": 0.001,
   "load_best_model_at_end": true,
@@ -75,6 +77,7 @@
     "use_bilateral_predicate_marker": true
   },
   "qanom_joint_factor": 14,
+  "relative_attention_max_distance": 128,
   "relative_attention_num_buckets": 32,
   "save_steps": 500,
   "save_strategy": "steps",
@@ -112,7 +115,7 @@
   "torch_dtype": "float32",
   "train_dataset": "joint_qanom",
   "train_epochs": 20,
-  "transformers_version": "4.14.1",
+  "transformers_version": "4.26.1",
   "use_bilateral_predicate_marker": true,
   "use_cache": true,
   "vocab_size": 32101

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "decoder_start_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.26.1"
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:67987af0aa8b7d2584e73cfaa33571bae84c52e4bc3534c5e896e1e7824343bd
-size 242028475
+oid sha256:dd5299f5b0ff532dfaf236df1d385a9f464c7163fee79f6c351410c96bc3554f
+size 242016842