初始化项目，由ModelHub XC社区提供模型

Model: dkleczek/papuGaPT2
Source: Original Platform

allegro_reviews/config.json

@@ -0,0 +1,36 @@
+{
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.0,
+  "bos_token_id": 50256,
+  "embd_pdrop": 0.0,
+  "eos_token_id": 50256,
+  "gradient_checkpointing": false,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 1024,
+  "n_embd": 768,
+  "n_head": 12,
+  "n_inner": null,
+  "n_layer": 12,
+  "n_positions": 1024,
+  "resid_pdrop": 0.0,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "transformers_version": "4.9.0.dev0",
+  "use_cache": true,
+  "vocab_size": 50257
+}

allegro_reviews/create_config_allegro.py

@@ -0,0 +1,6 @@
+from transformers import GPT2Config
+
+model_dir = "."  # ${MODEL_DIR}
+
+config = GPT2Config.from_pretrained("gpt2", resid_pdrop=0.0, embd_pdrop=0.0, attn_pdrop=0.0)
+config.save_pretrained(model_dir)

allegro_reviews/events.out.tfevents.1625481245.t1v-n-5d840006-w-0.20165.3.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25a5b7d6e069647cf953e1684211cf4b87049ae4e05610e37b1047966bd36fcc
+size 40

allegro_reviews/events.out.tfevents.1625482183.t1v-n-5d840006-w-0.22476.3.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee76cbdc38f6bec33ee28c5225264d95b8d46c0a2941ce59fbe8893f798a3de8
+size 40

allegro_reviews/events.out.tfevents.1625482418.t1v-n-5d840006-w-0.24291.3.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d20520f97baa97ebd08bbf9f66afb294613261a1661dbd9bf18ca39b4258e03d
+size 40

allegro_reviews/train_tokenizer_allegro.py

@@ -0,0 +1,26 @@
+from datasets import load_dataset
+from tokenizers import trainers, Tokenizer, normalizers, ByteLevelBPETokenizer
+
+model_dir = "."  # ${MODEL_DIR}
+
+# load dataset
+dataset = load_dataset("allegro_reviews", split="train")
+
+# Instantiate tokenizer
+tokenizer = ByteLevelBPETokenizer()
+
+def batch_iterator(batch_size=1000):
+    for i in range(0, len(dataset), batch_size):
+        yield dataset[i: i + batch_size]["text"]
+
+# Customized training
+tokenizer.train_from_iterator(batch_iterator(), vocab_size=50265, min_frequency=2, special_tokens=[
+    "<s>",
+    "<pad>",
+    "</s>",
+    "<unk>",
+    "<mask>",
+])
+
+# Save files to disk
+tokenizer.save(f"{model_dir}/tokenizer.json")