Changing to a less trained GPT-small

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "microsoft/DialoGPT-medium",
+  "_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": 1024,
-  "n_head": 16,
+  "n_embd": 768,
+  "n_head": 12,
   "n_inner": null,
-  "n_layer": 24,
+  "n_layer": 12,
   "n_positions": 1024,
   "resid_pdrop": 0.1,
   "scale_attn_weights": true,

eval_results.txt

@@ -1 +1 @@
-perplexity = tensor(4.2456)
+perplexity = tensor(7.6071)

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8a6b6de3faf2650b64408d34ab235d237b05155f6fbbb0e12031ddd2ee21faf4
-size 1444530764
+oid sha256:adc0a105f61321dba53dd3301745b6ddadb39ad32e39511859e137852c9eff79
+size 510378337

tokenizer_config.json

@@ -1 +1 @@
-{"unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "microsoft/DialoGPT-medium", "errors": "replace", "tokenizer_class": "GPT2Tokenizer"}
+{"unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "microsoft/DialoGPT-small", "errors": "replace", "tokenizer_class": "GPT2Tokenizer"}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6ae5414c5ca6649ce87c5f94c3714ce3217628bbfe91d62fc790a3d81965de43
-size 1077
+oid sha256:9bb4c7a39872e0b554db507498a5d91a912208989ed0e1921b8312384f54c87b
+size 1083