initial release

README.md

@@ -0,0 +1,27 @@
+---
+language:
+- "uk"
+tags:
+- "ukrainian"
+- "masked-lm"
+datasets:
+- "Goader/kobza"
+license: "apache-2.0"
+pipeline_tag: "fill-mask"
+mask_token: "<mask>"
+---
+
+# modernbert-base-ukrainian
+
+## Model Description
+
+This is a ModernBERT model pre-trained on Ukrainian texts. NVIDIA A100-SXM4-40GB×8 took 222 hours 58 minutes for training. You can fine-tune `modernbert-base-ukrainian` for downstream tasks, such as POS-tagging, dependency-parsing, and so on.
+
+## How to Use
+
+```py
+from transformers import AutoTokenizer,AutoModelForMaskedLM
+tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/modernbert-base-ukrainian")
+model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/modernbert-base-ukrainian")
+```
+

config.json

@@ -0,0 +1,46 @@
+{
+  "architectures": [
+    "ModernBertForMaskedLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "classifier_activation": "gelu",
+  "classifier_bias": false,
+  "classifier_dropout": 0.0,
+  "classifier_pooling": "mean",
+  "cls_token_id": 2,
+  "decoder_bias": true,
+  "deterministic_flash_attn": false,
+  "embedding_dropout": 0.0,
+  "eos_token_id": 3,
+  "global_attn_every_n_layers": 3,
+  "global_rope_theta": 160000.0,
+  "gradient_checkpointing": false,
+  "hidden_activation": "gelu",
+  "hidden_size": 768,
+  "initializer_cutoff_factor": 2.0,
+  "initializer_range": 0.02,
+  "intermediate_size": 1152,
+  "layer_norm_eps": 1e-05,
+  "local_attention": 128,
+  "local_rope_theta": 10000.0,
+  "max_position_embeddings": 8192,
+  "mlp_bias": false,
+  "mlp_dropout": 0.0,
+  "model_type": "modernbert",
+  "norm_bias": false,
+  "norm_eps": 1e-05,
+  "num_attention_heads": 12,
+  "num_hidden_layers": 22,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "repad_logits_with_grad": false,
+  "sep_token_id": 3,
+  "sparse_pred_ignore_index": -100,
+  "sparse_prediction": false,
+  "tokenizer_class": "LlamaTokenizerFast",
+  "torch_dtype": "float32",
+  "transformers_version": "4.54.0",
+  "vocab_size": 64000
+}

maker.py

@@ -0,0 +1,26 @@
+#! /usr/bin/python3
+tgt="KoichiYasuoka/modernbert-base-ukrainian"
+import os
+from transformers import LlamaTokenizerFast
+tkz=LlamaTokenizerFast.from_pretrained("KoichiYasuoka/modernbert-large-ukrainian-ud-embeds")
+tkz.save_pretrained(tgt)
+with open("train.py","w",encoding="utf-8") as w:
+  print(f'#! /usr/bin/env deepspeed\ntgt="{tgt}"'+'''
+from transformers import LlamaTokenizerFast,ModernBertForMaskedLM,AutoConfig,DataCollatorForLanguageModeling,TrainingArguments,Trainer
+tkz=LlamaTokenizerFast.from_pretrained(tgt)
+c={"vocab_size":len(tkz),"tokenizer_class":type(tkz).__name__}
+for k,v in tkz.special_tokens_map.items():
+  c[k+"_id"]=tkz.convert_tokens_to_ids(v)
+cfg=AutoConfig.from_pretrained("KoichiYasuoka/modernbert-base-classical-chinese",**c)
+arg=TrainingArguments(num_train_epochs=1,per_device_train_batch_size=1,output_dir=tgt,overwrite_output_dir=True,save_total_limit=2,save_safetensors=False)
+class trainTextDS(object):
+  def __init__(self,dataset,tokenizer):
+    from datasets import load_dataset
+    self.tokenizer=tokenizer
+    self.dataset=load_dataset(dataset)["train"]
+  __len__=lambda self:int(len(self.dataset)/32)
+  __getitem__=lambda self,i:self.tokenizer(" ".join(self.dataset[i*32:i*32+32]["text"]).replace("\\n"," "),truncation=True,add_special_tokens=True,max_length=8190)
+trn=Trainer(args=arg,data_collator=DataCollatorForLanguageModeling(tkz),model=ModernBertForMaskedLM(cfg),train_dataset=trainTextDS("Goader/kobza",tkz))
+trn.train()
+trn.save_model(tgt)''',file=w)
+os.system("chmod 755 train.py ; ./train.py")

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6950f3e3e5c6ba2013853ee8721ed27bcd8961fd29057496bf859e817b8a93d8
+size 640598126

special_tokens_map.json

@@ -0,0 +1,51 @@
+{
+  "bos_token": {
+    "content": "<cls>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "cls_token": {
+    "content": "<cls>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<sep>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "<sep>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c9699b255aa5ddd6575f1f3834454778153ebb60f957ac139d7b1685865e5e7
+size 2404944

tokenizer_config.json

@@ -0,0 +1,322 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": true,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<cls>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<sep>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<mask>",
+      "lstrip": true,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "!",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "6": {
+      "content": "\"",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "7": {
+      "content": "#",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "8": {
+      "content": "$",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "9": {
+      "content": "%",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "10": {
+      "content": "&",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "11": {
+      "content": "'",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "12": {
+      "content": "(",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "13": {
+      "content": ")",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "14": {
+      "content": "*",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "15": {
+      "content": "+",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "16": {
+      "content": ",",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "17": {
+      "content": "-",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "18": {
+      "content": ".",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "19": {
+      "content": "/",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "20": {
+      "content": ":",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "21": {
+      "content": ";",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "22": {
+      "content": "<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "23": {
+      "content": "=",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "24": {
+      "content": ">",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "25": {
+      "content": "?",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "26": {
+      "content": "@",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "27": {
+      "content": "[",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "28": {
+      "content": "\\",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "29": {
+      "content": "]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "30": {
+      "content": "^",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "31": {
+      "content": "_",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "32": {
+      "content": "`",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "33": {
+      "content": "{",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "34": {
+      "content": "|",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "35": {
+      "content": "}",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "36": {
+      "content": "~",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "bos_token": "<cls>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<cls>",
+  "eos_token": "<sep>",
+  "extra_special_tokens": {},
+  "legacy": true,
+  "mask_token": "<mask>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "sep_token": "<sep>",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}