Initial commit

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "microsoft/DialoGPT-large",
+  "_name_or_path": "microsoft/DialoGPT-small",
   "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"
@@ -12,10 +12,10 @@
   "layer_norm_epsilon": 1e-05,
   "model_type": "gpt2",
   "n_ctx": 1024,
-  "n_embd": 1280,
-  "n_head": 20,
+  "n_embd": 768,
+  "n_head": 12,
   "n_inner": null,
-  "n_layer": 36,
+  "n_layer": 12,
   "n_positions": 1024,
   "reorder_and_upcast_attn": false,
   "resid_pdrop": 0.1,

eval_results.txt

@@ -1 +1 @@
-perplexity = tensor(1.1794)
+perplexity = tensor(2.1727)

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35883bc17ecdad0b44964eccc285f3ce6bf1d0b6639562e1d2d23d54e2b91373
-size 3134035005
+oid sha256:d69ebe3595a4e3cd0e59f8815e496c3667a5bd9fd758289037e7aa6b8f603bc1
+size 510398013

tokenizer_config.json

@@ -19,7 +19,7 @@
   },
   "errors": "replace",
   "model_max_length": 1024,
-  "name_or_path": "microsoft/DialoGPT-large",
+  "name_or_path": "microsoft/DialoGPT-small",
   "pad_token": null,
   "special_tokens_map_file": null,
   "tokenizer_class": "GPT2Tokenizer",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a18ec590b152442d93d60caf042ea4f4bb3e98803cbeb42d8913cb3d9f1a2be
+oid sha256:3915c50cfa279e308c4199843fae1675730cc746765854db5abf008c262db96d
 size 1339