up

config.json

@@ -7,26 +7,20 @@
   "attention_dropout": 0.0,
   "bos_token_id": 0,
   "d_model": 7168,
-  "decoder_layernorm": false,
-  "decoder_start_token_id": 2,
   "do_layer_norm_before": true,
   "dropout": 0.1,
   "eos_token_id": 2,
   "ffn_dim": 28672,
-  "forced_eos_token_id": 2,
   "init_std": 0.02,
   "layerdrop": 0.0,
   "max_position_embeddings": 2048,
   "model_type": "opt",
   "num_attention_heads": 56,
   "num_hidden_layers": 48,
-  "output_projection": true,
   "pad_token_id": 1,
-  "scale_embedding": false,
-  "share_input_output_embed": true,
   "torch_dtype": "float16",
   "transformers_version": "4.19.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 50272,
   "word_embed_proj_dim": 7168
 }