mBART + fine-tuned dbmdz/german-gpt2

config.json

@@ -49,16 +49,16 @@
   "d_model": 1024,
   "decoder_attention_heads": 16,
   "decoder_config": {
-    "_name_or_path": "benjamin/gerpt2",
+    "_name_or_path": "josh-oo/german-gpt2-easy",
     "activation_function": "gelu_new",
-    "add_cross_attention": false,
+    "add_cross_attention": true,
     "architectures": [
       "GPT2LMHeadModel"
     ],
     "attn_pdrop": 0.1,
     "bad_words_ids": null,
     "begin_suppress_tokens": null,
-    "bos_token_id": 0,
+    "bos_token_id": 50267,
     "chunk_size_feed_forward": 0,
     "cross_attention_hidden_size": null,
     "decoder_start_token_id": null,
@@ -67,7 +67,7 @@
     "early_stopping": false,
     "embd_pdrop": 0.1,
     "encoder_no_repeat_ngram_size": 0,
-    "eos_token_id": 0,
+    "eos_token_id": 50266,
     "exponential_decay_length_penalty": null,
     "finetuning_task": null,
     "forced_bos_token_id": null,
@@ -78,7 +78,7 @@
       "1": "LABEL_1"
     },
     "initializer_range": 0.02,
-    "is_decoder": false,
+    "is_decoder": true,
     "is_encoder_decoder": false,
     "label2id": {
       "LABEL_0": 0,
@@ -102,7 +102,7 @@
     "output_attentions": false,
     "output_hidden_states": false,
     "output_scores": false,
-    "pad_token_id": 1,
+    "pad_token_id": 50268,
     "prefix": null,
     "problem_type": null,
     "pruned_heads": {},
@@ -124,13 +124,13 @@
     "task_specific_params": {
       "text-generation": {
         "do_sample": true,
-        "max_length": 100
+        "max_length": 50
       }
     },
     "temperature": 1.0,
     "tf_legacy_loss": false,
     "tie_encoder_decoder": false,
-    "tie_word_embeddings": false,
+    "tie_word_embeddings": true,
     "tokenizer_class": null,
     "top_k": 50,
     "top_p": 1.0,
@@ -140,7 +140,7 @@
     "typical_p": 1.0,
     "use_bfloat16": false,
     "use_cache": true,
-    "vocab_size": 50258
+    "vocab_size": 50269
   },
   "decoder_ffn_dim": 4096,
   "decoder_layerdrop": 0.0,
@@ -150,7 +150,7 @@
   "encoder_ffn_dim": 4096,
   "encoder_layerdrop": 0.0,
   "encoder_layers": 12,
-  "eos_token_id": 0,
+  "eos_token_id": 50266,
   "forced_eos_token_id": 2,
   "from_mbart": false,
   "global_attention_indices": [

generation_config.json

@@ -1,7 +1,7 @@
 {
   "_from_model_config": true,
   "bos_token_id": 0,
-  "eos_token_id": 0,
+  "eos_token_id": 50266,
   "forced_eos_token_id": 2,
   "max_length": 1024,
   "num_beams": 5,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:74434cbc4348f9491b766c9acc11ac4a55a46e00f6e61d486d77e41094a386c1
-size 1648941221
+oid sha256:c748af332639939d651577ce0c95204a617fab747d23854ddd6bd809b475685f
+size 1649008805