Upload 9 files

config.json

@@ -0,0 +1,47 @@
+{
+  "_name_or_path": "/data/hypertext/xpkong/newcode/checkpoints/whr-vary-qwen-sft-1-8-1-9-3/",
+  "architectures": [
+    "MMGPTQwenForCausalLM"
+  ],
+  "attn_dropout_prob": 0.0,
+  "auto_map": {
+    "AutoConfig": "configuration_qwen.QWenConfig",
+    "AutoModelForCausalLM": "modeling_qwen.QWenLMHeadModel"
+  },
+  "bf16": false,
+  "emb_dropout_prob": 0.0,
+  "fp16": false,
+  "fp32": false,
+  "freeze_vision_tower": false,
+  "hidden_size": 2048,
+  "im_end_token": 151858,
+  "im_patch_token": 151859,
+  "im_start_token": 151857,
+  "image_token_len": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "kv_channels": 128,
+  "layer_norm_epsilon": 1e-06,
+  "max_position_embeddings": 8192,
+  "model_type": "mmgpt",
+  "no_bias": true,
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "onnx_safe": null,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 1.0,
+  "scale_attn_weights": true,
+  "seq_length": 8192,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "QWenTokenizer",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.32.1",
+  "use_cache": true,
+  "use_dynamic_ntk": true,
+  "use_flash_attn": "auto",
+  "use_im_start_end": true,
+  "use_logn_attn": true,
+  "vision_select_layer": -2,
+  "vision_tower": "/mnt/host0/vit-large-patch14",
+  "vocab_size": 151860
+}

generation_config.json

@@ -0,0 +1,12 @@
+{
+  "chat_format": "chatml",
+  "do_sample": true,
+  "eos_token_id": 151643,
+  "max_new_tokens": 512,
+  "max_window_size": 6144,
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "top_k": 0,
+  "top_p": 0.8,
+  "transformers_version": "4.32.1"
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbb621547e4231108139d53ca94c401e71cf7767f80931da1beeb6ddd5b3f129
+size 4474993293

special_tokens_map.json

@@ -0,0 +1,3 @@
+{
+  "pad_token": "<|endoftext|>"
+}

tokenization_qwen.py

@@ -0,0 +1,264 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Tokenization classes for QWen."""
+
+import base64
+import logging
+import os
+import unicodedata
+from typing import Collection, Dict, List, Set, Tuple, Union
+
+import tiktoken
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+
+VOCAB_FILES_NAMES = {"vocab_file": "qwen.tiktoken"}
+
+PAT_STR = r"""(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"""
+ENDOFTEXT = "<|endoftext|>"
+IMSTART = "<|im_start|>"
+IMEND = "<|im_end|>"
+# as the default behavior is changed to allow special tokens in
+# regular texts, the surface forms of special tokens need to be
+# as different as possible to minimize the impact
+EXTRAS = tuple((f"<|extra_{i}|>" for i in range(205)))
+SPECIAL_TOKENS = (
+    ENDOFTEXT,
+    IMSTART,
+    IMEND,
+) + EXTRAS
+
+
+def _load_tiktoken_bpe(tiktoken_bpe_file: str) -> Dict[bytes, int]:
+    with open(tiktoken_bpe_file, "rb") as f:
+        contents = f.read()
+    return {
+        base64.b64decode(token): int(rank)
+        for token, rank in (line.split() for line in contents.splitlines() if line)
+    }
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    vocab_files_names = VOCAB_FILES_NAMES
+
+    def __init__(
+        self,
+        vocab_file,
+        errors="replace",
+        image_start_tag='<img>',
+        image_end_tag='</img>',
+        image_pad_tag='<imgpad>',
+        ref_start_tag='<ref>',
+        ref_end_tag='</ref>',
+        box_start_tag='<box>',
+        box_end_tag='</box>',
+        quad_start_tag='<quad>',
+        quad_end_tag='</quad>',
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        
+        self.image_start_tag = image_start_tag
+        self.image_end_tag = image_end_tag
+        self.image_pad_tag = image_pad_tag
+        self.ref_start_tag = ref_start_tag
+        self.ref_end_tag = ref_end_tag
+        self.box_start_tag = box_start_tag
+        self.box_end_tag = box_end_tag
+        self.quad_start_tag = quad_start_tag
+        self.quad_end_tag = quad_end_tag
+        self.IMAGE_ST = (
+            ref_start_tag, ref_end_tag,
+            box_start_tag, box_end_tag,
+            quad_start_tag, quad_end_tag,
+            image_start_tag, image_end_tag,
+            image_pad_tag
+        )
+
+        self.errors = errors  # how to handle errors in decoding
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in enumerate(
+                SPECIAL_TOKENS + self.IMAGE_ST, start=len(self.mergeable_ranks)
+            )
+        }
+        
+        self.img_start_id = self.special_tokens[self.image_start_tag]
+        self.img_end_id = self.special_tokens[self.image_end_tag]
+        self.img_pad_id = self.special_tokens[self.image_pad_tag]
+        self.ref_start_id = self.special_tokens[self.ref_start_tag]
+        self.ref_end_id = self.special_tokens[self.ref_end_tag]
+        self.box_start_id = self.special_tokens[self.box_start_tag]
+        self.box_end_id = self.special_tokens[self.box_end_tag]
+        self.quad_start_id = self.special_tokens[self.quad_start_tag]
+        self.quad_end_id = self.special_tokens[self.quad_end_tag]
+
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        assert (
+            len(self.mergeable_ranks) + len(self.special_tokens) == enc.n_vocab
+        ), f"{len(self.mergeable_ranks) + len(self.special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.decoder = {
+            v: k for k, v in self.mergeable_ranks.items()
+        }  # type: dict[int, bytes|str]
+        self.decoder.update({v: k for k, v in self.special_tokens.items()})
+
+        self.tokenizer = enc  # type: tiktoken.Encoding
+
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = self.special_tokens[IMSTART]
+        self.im_end_id = self.special_tokens[IMEND]
+
+    def __len__(self) -> int:
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self) -> Dict[bytes, int]:
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(
+        self, tokens: Union[bytes, str, List[Union[bytes, str]]]
+    ) -> List[int]:
+        ids = []
+        if isinstance(tokens, (str, bytes)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.mergeable_ranks.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.mergeable_ranks.get(token))
+        return ids
+
+    def _add_tokens(self, new_tokens: Union[List[str], List[AddedToken]], special_tokens: bool = False) -> int:
+        if not special_tokens and new_tokens:
+            raise ValueError('Adding regular tokens is not supported')
+        for token in new_tokens:
+            surface_form = token.content if isinstance(token, AddedToken) else token
+            if surface_form not in SPECIAL_TOKENS:
+                raise ValueError('Adding unknown special tokens is not supported')
+        return 0
+
+    def save_vocabulary(self, save_directory: str, **kwargs) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary).
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        file_path = os.path.join(save_directory, "qwen.tiktoken")
+        with open(file_path, "w", encoding="utf8") as w:
+            for k, v in self.mergeable_ranks.items():
+                line = base64.b64encode(k).decode("utf8") + " " + str(v) + "\n"
+                w.write(line)
+        return (file_path,)
+
+    def tokenize(
+        self,
+        text: str,
+        allowed_special: Union[Set, str] = "all",
+        disallowed_special: Union[Collection, str] = (),
+        **kwargs,
+    ) -> List[Union[bytes, str]]:
+        """
+        Converts a string in a sequence of tokens.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            allowed_special (`Literal["all"]` or `set`):
+                The surface forms of the tokens to be encoded as special tokens in regular texts.
+                Default to "all".
+            disallowed_special (`Literal["all"]` or `Collection`):
+                The surface forms of the tokens that should not be in regular texts and trigger errors.
+                Default to an empty tuple.
+
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method.
+
+        Returns:
+            `List[bytes|str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+
+        # this implementation takes a detour: text -> token id -> token surface forms
+        for t in self.tokenizer.encode(
+            text, allowed_special=allowed_special, disallowed_special=disallowed_special
+        ):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[Union[bytes, str]]) -> str:
+        """
+        Converts a sequence of tokens in a single string.
+        """
+        text = ""
+        temp = b""
+        for t in tokens:
+            if isinstance(t, str):
+                if temp:
+                    text += temp.decode("utf-8", errors=self.errors)
+                    temp = b""
+                text += t
+            elif isinstance(t, bytes):
+                temp += t
+            else:
+                raise TypeError("token should only be of type types or str")
+        if temp:
+            text += temp.decode("utf-8", errors=self.errors)
+        return text
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.n_vocab
+
+    def _convert_id_to_token(self, index: int) -> Union[bytes, str]:
+        """Converts an id to a token, special tokens included"""
+        if index in self.decoder:
+            return self.decoder[index]
+        raise ValueError("unknown ids")
+
+    def _convert_token_to_id(self, token: Union[bytes, str]) -> int:
+        """Converts a token to an id using the vocab, special tokens included"""
+        if token in self.special_tokens:
+            return self.special_tokens[token]
+        if token in self.mergeable_ranks:
+            return self.mergeable_ranks[token]
+        raise ValueError("unknown token")
+
+    def _tokenize(self, text: str, **kwargs):
+        """
+        Converts a string in a sequence of tokens (string), using the tokenizer. Split in words for word-based
+        vocabulary or sub-words for sub-word-based vocabularies (BPE/SentencePieces/WordPieces).
+
+        Do NOT take care of added tokens.
+        """
+        raise NotImplementedError
+
+    def _decode(
+        self,
+        token_ids: Union[int, List[int]],
+        skip_special_tokens: bool = False,
+        errors: str = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        if skip_special_tokens:
+            token_ids = [i for i in token_ids if i < self.eod_id]
+        return self.tokenizer.decode(token_ids, errors=errors or self.errors)

tokenizer_config.json

@@ -0,0 +1,12 @@
+{
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+    ]
+  },
+  "clean_up_tokenization_spaces": true,
+  "model_max_length": 4096,
+  "padding_side": "right",
+  "tokenizer_class": "QWenTokenizer"
+}

trainer_state.json

@@ -0,0 +1,2584 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 426,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.5384615384615387e-06,
+      "loss": 0.92,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.0769230769230774e-06,
+      "loss": 0.9533,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.615384615384616e-06,
+      "loss": 0.9873,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.153846153846155e-06,
+      "loss": 0.9105,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.692307692307694e-06,
+      "loss": 0.9502,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.230769230769232e-06,
+      "loss": 0.9663,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.076923076923077e-05,
+      "loss": 0.9044,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.230769230769231e-05,
+      "loss": 0.9281,
+      "step": 8
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3846153846153847e-05,
+      "loss": 0.8729,
+      "step": 9
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5384615384615387e-05,
+      "loss": 0.894,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6923076923076924e-05,
+      "loss": 0.9235,
+      "step": 11
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8461538461538465e-05,
+      "loss": 0.905,
+      "step": 12
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2e-05,
+      "loss": 0.908,
+      "step": 13
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999710687264073e-05,
+      "loss": 0.9524,
+      "step": 14
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999884276579666e-05,
+      "loss": 0.9495,
+      "step": 15
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9997396285817907e-05,
+      "loss": 0.8929,
+      "step": 16
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9995371331024835e-05,
+      "loss": 0.9357,
+      "step": 17
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999276801858648e-05,
+      "loss": 0.9182,
+      "step": 18
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9989586499137137e-05,
+      "loss": 0.9526,
+      "step": 19
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.998582695676762e-05,
+      "loss": 0.9119,
+      "step": 20
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.998148960901463e-05,
+      "loss": 0.9072,
+      "step": 21
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9976574706848154e-05,
+      "loss": 0.888,
+      "step": 22
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9971082534656958e-05,
+      "loss": 0.9345,
+      "step": 23
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.99650134102321e-05,
+      "loss": 0.954,
+      "step": 24
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995836768474859e-05,
+      "loss": 0.9353,
+      "step": 25
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9951145742745027e-05,
+      "loss": 0.8974,
+      "step": 26
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9943348002101374e-05,
+      "loss": 0.9084,
+      "step": 27
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9934974914014765e-05,
+      "loss": 0.9109,
+      "step": 28
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9926026962973407e-05,
+      "loss": 0.9312,
+      "step": 29
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9916504666728533e-05,
+      "loss": 0.8948,
+      "step": 30
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9906408576264468e-05,
+      "loss": 0.8773,
+      "step": 31
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9895739275766717e-05,
+      "loss": 0.9831,
+      "step": 32
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9884497382588185e-05,
+      "loss": 0.9739,
+      "step": 33
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9872683547213446e-05,
+      "loss": 0.9398,
+      "step": 34
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.986029845322111e-05,
+      "loss": 0.8877,
+      "step": 35
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9847342817244256e-05,
+      "loss": 0.8749,
+      "step": 36
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9833817388928987e-05,
+      "loss": 0.9456,
+      "step": 37
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9819722950891034e-05,
+      "loss": 0.9641,
+      "step": 38
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.980506031867049e-05,
+      "loss": 0.8627,
+      "step": 39
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9789830340684593e-05,
+      "loss": 0.941,
+      "step": 40
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9774033898178668e-05,
+      "loss": 0.9602,
+      "step": 41
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9757671905175118e-05,
+      "loss": 0.9361,
+      "step": 42
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.974074530842053e-05,
+      "loss": 0.8819,
+      "step": 43
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.972325508733091e-05,
+      "loss": 0.9098,
+      "step": 44
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9705202253935004e-05,
+      "loss": 0.974,
+      "step": 45
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9686587852815733e-05,
+      "loss": 0.9917,
+      "step": 46
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9667412961049755e-05,
+      "loss": 0.9431,
+      "step": 47
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9647678688145163e-05,
+      "loss": 0.8762,
+      "step": 48
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.962738617597724e-05,
+      "loss": 0.8853,
+      "step": 49
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9606536598722435e-05,
+      "loss": 0.8557,
+      "step": 50
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.95851311627904e-05,
+      "loss": 0.8813,
+      "step": 51
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.956317110675417e-05,
+      "loss": 0.9625,
+      "step": 52
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9540657701278536e-05,
+      "loss": 0.9091,
+      "step": 53
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9517592249046476e-05,
+      "loss": 0.9151,
+      "step": 54
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9493976084683814e-05,
+      "loss": 0.9429,
+      "step": 55
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.946981057468197e-05,
+      "loss": 0.8978,
+      "step": 56
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9445097117318915e-05,
+      "loss": 0.9288,
+      "step": 57
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9419837142578228e-05,
+      "loss": 0.9247,
+      "step": 58
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9394032112066405e-05,
+      "loss": 0.8868,
+      "step": 59
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9367683518928228e-05,
+      "loss": 0.9086,
+      "step": 60
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9340792887760417e-05,
+      "loss": 0.8921,
+      "step": 61
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9313361774523387e-05,
+      "loss": 0.8934,
+      "step": 62
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.928539176645122e-05,
+      "loss": 0.8898,
+      "step": 63
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.925688448195983e-05,
+      "loss": 0.9609,
+      "step": 64
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.922784157055331e-05,
+      "loss": 0.931,
+      "step": 65
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9198264712728494e-05,
+      "loss": 0.9022,
+      "step": 66
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9168155619877712e-05,
+      "loss": 0.8409,
+      "step": 67
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9137516034189768e-05,
+      "loss": 0.9554,
+      "step": 68
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9106347728549134e-05,
+      "loss": 0.8817,
+      "step": 69
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.907465250643337e-05,
+      "loss": 0.9268,
+      "step": 70
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9042432201808757e-05,
+      "loss": 0.9427,
+      "step": 71
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.900968867902419e-05,
+      "loss": 0.9137,
+      "step": 72
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.897642383270331e-05,
+      "loss": 0.8941,
+      "step": 73
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8942639587634854e-05,
+      "loss": 0.8784,
+      "step": 74
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8908337898661287e-05,
+      "loss": 0.9041,
+      "step": 75
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8873520750565716e-05,
+      "loss": 0.8818,
+      "step": 76
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8838190157957013e-05,
+      "loss": 0.8648,
+      "step": 77
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.880234816515326e-05,
+      "loss": 0.9294,
+      "step": 78
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8765996846063456e-05,
+      "loss": 0.9423,
+      "step": 79
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8729138304067512e-05,
+      "loss": 0.9467,
+      "step": 80
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8691774671894562e-05,
+      "loss": 0.8737,
+      "step": 81
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8653908111499534e-05,
+      "loss": 0.8909,
+      "step": 82
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8615540813938063e-05,
+      "loss": 0.912,
+      "step": 83
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8576674999239713e-05,
+      "loss": 0.9153,
+      "step": 84
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8537312916279526e-05,
+      "loss": 0.9053,
+      "step": 85
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.849745684264788e-05,
+      "loss": 0.9688,
+      "step": 86
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.845710908451872e-05,
+      "loss": 0.9658,
+      "step": 87
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.84162719765161e-05,
+      "loss": 0.9626,
+      "step": 88
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8374947881579116e-05,
+      "loss": 0.9328,
+      "step": 89
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.833313919082515e-05,
+      "loss": 0.9245,
+      "step": 90
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8290848323411554e-05,
+      "loss": 0.8959,
+      "step": 91
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8248077726395634e-05,
+      "loss": 0.8694,
+      "step": 92
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8204829874593083e-05,
+      "loss": 0.8948,
+      "step": 93
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8161107270434757e-05,
+      "loss": 0.9032,
+      "step": 94
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8116912443821912e-05,
+      "loss": 0.8311,
+      "step": 95
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8072247951979785e-05,
+      "loss": 0.9178,
+      "step": 96
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8027116379309637e-05,
+      "loss": 0.9193,
+      "step": 97
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.798152033723923e-05,
+      "loss": 0.8684,
+      "step": 98
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7935462464071697e-05,
+      "loss": 0.8673,
+      "step": 99
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7888945424832896e-05,
+      "loss": 0.8611,
+      "step": 100
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.78419719111172e-05,
+      "loss": 0.9236,
+      "step": 101
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7794544640931773e-05,
+      "loss": 0.8275,
+      "step": 102
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.774666635853927e-05,
+      "loss": 0.867,
+      "step": 103
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7698339834299064e-05,
+      "loss": 0.9036,
+      "step": 104
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7649567864506943e-05,
+      "loss": 0.9314,
+      "step": 105
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7600353271233316e-05,
+      "loss": 0.8757,
+      "step": 106
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7550698902159897e-05,
+      "loss": 0.9196,
+      "step": 107
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7500607630414973e-05,
+      "loss": 0.8508,
+      "step": 108
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.745008235440711e-05,
+      "loss": 0.8699,
+      "step": 109
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7399125997657478e-05,
+      "loss": 0.8734,
+      "step": 110
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7347741508630673e-05,
+      "loss": 0.8488,
+      "step": 111
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.729593186056412e-05,
+      "loss": 0.8975,
+      "step": 112
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7243700051296016e-05,
+      "loss": 0.9325,
+      "step": 113
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.71910491030919e-05,
+      "loss": 0.8745,
+      "step": 114
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.713798206246974e-05,
+      "loss": 0.945,
+      "step": 115
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.708450200002368e-05,
+      "loss": 0.9282,
+      "step": 116
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.703061201024636e-05,
+      "loss": 0.8688,
+      "step": 117
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6976315211349848e-05,
+      "loss": 0.8962,
+      "step": 118
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6921614745085236e-05,
+      "loss": 0.8591,
+      "step": 119
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6866513776560844e-05,
+      "loss": 0.8918,
+      "step": 120
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6811015494059048e-05,
+      "loss": 0.9125,
+      "step": 121
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6755123108851846e-05,
+      "loss": 0.8773,
+      "step": 122
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.669883985501501e-05,
+      "loss": 0.9334,
+      "step": 123
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6642168989240977e-05,
+      "loss": 0.887,
+      "step": 124
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.658511379065039e-05,
+      "loss": 0.8577,
+      "step": 125
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6527677560602364e-05,
+      "loss": 0.8751,
+      "step": 126
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6469863622503478e-05,
+      "loss": 0.8974,
+      "step": 127
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.641167532161545e-05,
+      "loss": 0.9063,
+      "step": 128
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6353116024861584e-05,
+      "loss": 0.8761,
+      "step": 129
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6294189120631956e-05,
+      "loss": 0.8967,
+      "step": 130
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6234898018587336e-05,
+      "loss": 0.8966,
+      "step": 131
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.617524614946192e-05,
+      "loss": 0.901,
+      "step": 132
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6115236964864798e-05,
+      "loss": 0.843,
+      "step": 133
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6054873937080248e-05,
+      "loss": 0.8904,
+      "step": 134
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5994160558866803e-05,
+      "loss": 0.9368,
+      "step": 135
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5933100343255184e-05,
+      "loss": 0.9121,
+      "step": 136
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5871696823344998e-05,
+      "loss": 0.928,
+      "step": 137
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.580995355210031e-05,
+      "loss": 0.8595,
+      "step": 138
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5747874102144073e-05,
+      "loss": 0.827,
+      "step": 139
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5685462065551375e-05,
+      "loss": 0.8894,
+      "step": 140
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.562272105364164e-05,
+      "loss": 0.8695,
+      "step": 141
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5559654696769628e-05,
+      "loss": 0.8895,
+      "step": 142
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5496266644115386e-05,
+      "loss": 0.9051,
+      "step": 143
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5432560563473106e-05,
+      "loss": 0.8599,
+      "step": 144
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.536854014103888e-05,
+      "loss": 0.8624,
+      "step": 145
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5304209081197425e-05,
+      "loss": 0.8795,
+      "step": 146
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5239571106307729e-05,
+      "loss": 0.8165,
+      "step": 147
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5174629956487659e-05,
+      "loss": 0.8826,
+      "step": 148
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5109389389397566e-05,
+      "loss": 0.8126,
+      "step": 149
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5043853180022838e-05,
+      "loss": 0.8531,
+      "step": 150
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.4978025120455484e-05,
+      "loss": 0.8569,
+      "step": 151
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4911909019674703e-05,
+      "loss": 0.8385,
+      "step": 152
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4845508703326504e-05,
+      "loss": 0.8987,
+      "step": 153
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4778828013502318e-05,
+      "loss": 0.8636,
+      "step": 154
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4711870808516708e-05,
+      "loss": 0.8986,
+      "step": 155
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4644640962684107e-05,
+      "loss": 0.8355,
+      "step": 156
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4577142366094643e-05,
+      "loss": 0.8899,
+      "step": 157
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4509378924389044e-05,
+      "loss": 0.8985,
+      "step": 158
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4441354558532654e-05,
+      "loss": 0.8431,
+      "step": 159
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4373073204588556e-05,
+      "loss": 0.8805,
+      "step": 160
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4304538813489808e-05,
+      "loss": 0.8278,
+      "step": 161
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4235755350810854e-05,
+      "loss": 0.8554,
+      "step": 162
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4166726796538044e-05,
+      "loss": 0.8091,
+      "step": 163
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4097457144839362e-05,
+      "loss": 0.8755,
+      "step": 164
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4027950403833295e-05,
+      "loss": 0.8718,
+      "step": 165
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3958210595356924e-05,
+      "loss": 0.8415,
+      "step": 166
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.388824175473321e-05,
+      "loss": 0.8597,
+      "step": 167
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3818047930537491e-05,
+      "loss": 0.9381,
+      "step": 168
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3747633184363234e-05,
+      "loss": 0.8926,
+      "step": 169
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3677001590587011e-05,
+      "loss": 0.8472,
+      "step": 170
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3606157236132756e-05,
+      "loss": 0.8711,
+      "step": 171
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3535104220235262e-05,
+      "loss": 0.8187,
+      "step": 172
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3463846654203021e-05,
+      "loss": 0.8371,
+      "step": 173
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3392388661180303e-05,
+      "loss": 0.8185,
+      "step": 174
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3320734375908608e-05,
+      "loss": 0.8315,
+      "step": 175
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3248887944487397e-05,
+      "loss": 0.8923,
+      "step": 176
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3176853524134198e-05,
+      "loss": 0.8946,
+      "step": 177
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3104635282944054e-05,
+      "loss": 0.8792,
+      "step": 178
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3032237399648357e-05,
+      "loss": 0.8519,
+      "step": 179
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2959664063373044e-05,
+      "loss": 0.8592,
+      "step": 180
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2886919473396212e-05,
+      "loss": 0.8558,
+      "step": 181
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.281400783890513e-05,
+      "loss": 0.8826,
+      "step": 182
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2740933378752685e-05,
+      "loss": 0.8083,
+      "step": 183
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2667700321213282e-05,
+      "loss": 0.863,
+      "step": 184
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2594312903738164e-05,
+      "loss": 0.8179,
+      "step": 185
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2520775372710242e-05,
+      "loss": 0.785,
+      "step": 186
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2447091983198368e-05,
+      "loss": 0.8361,
+      "step": 187
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2373266998711152e-05,
+      "loss": 0.8412,
+      "step": 188
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2299304690950235e-05,
+      "loss": 0.8854,
+      "step": 189
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2225209339563144e-05,
+      "loss": 0.8221,
+      "step": 190
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2150985231895647e-05,
+      "loss": 0.8823,
+      "step": 191
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2076636662743673e-05,
+      "loss": 0.905,
+      "step": 192
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2002167934104815e-05,
+      "loss": 0.9039,
+      "step": 193
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1927583354929393e-05,
+      "loss": 0.8522,
+      "step": 194
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1852887240871145e-05,
+      "loss": 0.8775,
+      "step": 195
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.177808391403749e-05,
+      "loss": 0.8461,
+      "step": 196
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.170317770273946e-05,
+      "loss": 0.8631,
+      "step": 197
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.162817294124124e-05,
+      "loss": 0.8432,
+      "step": 198
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1553073969509382e-05,
+      "loss": 0.8647,
+      "step": 199
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1477885132961679e-05,
+      "loss": 0.815,
+      "step": 200
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1402610782215735e-05,
+      "loss": 0.8566,
+      "step": 201
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1327255272837221e-05,
+      "loss": 0.8163,
+      "step": 202
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1251822965087856e-05,
+      "loss": 0.8177,
+      "step": 203
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1176318223673106e-05,
+      "loss": 0.9081,
+      "step": 204
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1100745417489631e-05,
+      "loss": 0.8451,
+      "step": 205
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1025108919372501e-05,
+      "loss": 0.8246,
+      "step": 206
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0949413105842148e-05,
+      "loss": 0.8439,
+      "step": 207
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0873662356851164e-05,
+      "loss": 0.8444,
+      "step": 208
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0797861055530832e-05,
+      "loss": 0.852,
+      "step": 209
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0722013587937528e-05,
+      "loss": 0.8598,
+      "step": 210
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.064612434279892e-05,
+      "loss": 0.896,
+      "step": 211
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.057019771126004e-05,
+      "loss": 0.7676,
+      "step": 212
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0494238086629184e-05,
+      "loss": 0.8303,
+      "step": 213
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0418249864123724e-05,
+      "loss": 0.8253,
+      "step": 214
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0342237440615765e-05,
+      "loss": 0.8391,
+      "step": 215
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.026620521437775e-05,
+      "loss": 0.8656,
+      "step": 216
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0190157584827952e-05,
+      "loss": 0.8424,
+      "step": 217
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0114098952275935e-05,
+      "loss": 0.8154,
+      "step": 218
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0038033717667907e-05,
+      "loss": 0.8372,
+      "step": 219
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.961966282332094e-06,
+      "loss": 0.8529,
+      "step": 220
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.885901047724066e-06,
+      "loss": 0.8619,
+      "step": 221
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.80984241517205e-06,
+      "loss": 0.7912,
+      "step": 222
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.733794785622254e-06,
+      "loss": 0.8761,
+      "step": 223
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.657762559384237e-06,
+      "loss": 0.8112,
+      "step": 224
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.581750135876277e-06,
+      "loss": 0.8028,
+      "step": 225
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.505761913370814e-06,
+      "loss": 0.8751,
+      "step": 226
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.429802288739963e-06,
+      "loss": 0.8743,
+      "step": 227
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.353875657201084e-06,
+      "loss": 0.8358,
+      "step": 228
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.277986412062476e-06,
+      "loss": 0.7817,
+      "step": 229
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.202138944469168e-06,
+      "loss": 0.8076,
+      "step": 230
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.12633764314884e-06,
+      "loss": 0.8662,
+      "step": 231
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.050586894157853e-06,
+      "loss": 0.8079,
+      "step": 232
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.974891080627504e-06,
+      "loss": 0.8166,
+      "step": 233
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.899254582510369e-06,
+      "loss": 0.8975,
+      "step": 234
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.823681776326899e-06,
+      "loss": 0.7989,
+      "step": 235
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.748177034912148e-06,
+      "loss": 0.8004,
+      "step": 236
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.672744727162782e-06,
+      "loss": 0.8065,
+      "step": 237
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.597389217784268e-06,
+      "loss": 0.799,
+      "step": 238
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.522114867038323e-06,
+      "loss": 0.8044,
+      "step": 239
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.446926030490623e-06,
+      "loss": 0.7698,
+      "step": 240
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.371827058758763e-06,
+      "loss": 0.8325,
+      "step": 241
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.296822297260541e-06,
+      "loss": 0.8511,
+      "step": 242
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.221916085962511e-06,
+      "loss": 0.8323,
+      "step": 243
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.147112759128859e-06,
+      "loss": 0.8127,
+      "step": 244
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.072416645070609e-06,
+      "loss": 0.8726,
+      "step": 245
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.997832065895188e-06,
+      "loss": 0.824,
+      "step": 246
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.923363337256327e-06,
+      "loss": 0.7932,
+      "step": 247
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.849014768104354e-06,
+      "loss": 0.8338,
+      "step": 248
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.774790660436857e-06,
+      "loss": 0.8015,
+      "step": 249
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.700695309049768e-06,
+      "loss": 0.7834,
+      "step": 250
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.626733001288852e-06,
+      "loss": 0.8021,
+      "step": 251
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.552908016801633e-06,
+      "loss": 0.8361,
+      "step": 252
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.479224627289765e-06,
+      "loss": 0.8649,
+      "step": 253
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.40568709626184e-06,
+      "loss": 0.8319,
+      "step": 254
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.3322996787867225e-06,
+      "loss": 0.8105,
+      "step": 255
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.259066621247316e-06,
+      "loss": 0.878,
+      "step": 256
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.185992161094877e-06,
+      "loss": 0.7354,
+      "step": 257
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.113080526603793e-06,
+      "loss": 0.8544,
+      "step": 258
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.040335936626959e-06,
+      "loss": 0.8877,
+      "step": 259
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.967762600351646e-06,
+      "loss": 0.8115,
+      "step": 260
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.895364717055949e-06,
+      "loss": 0.8131,
+      "step": 261
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.8231464758658075e-06,
+      "loss": 0.8128,
+      "step": 262
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.7511120555126055e-06,
+      "loss": 0.8055,
+      "step": 263
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.6792656240913936e-06,
+      "loss": 0.8022,
+      "step": 264
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.607611338819697e-06,
+      "loss": 0.8304,
+      "step": 265
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.536153345796984e-06,
+      "loss": 0.8399,
+      "step": 266
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.464895779764742e-06,
+      "loss": 0.7836,
+      "step": 267
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.393842763867248e-06,
+      "loss": 0.7966,
+      "step": 268
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.3229984094129895e-06,
+      "loss": 0.8371,
+      "step": 269
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.252366815636768e-06,
+      "loss": 0.8918,
+      "step": 270
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.181952069462514e-06,
+      "loss": 0.819,
+      "step": 271
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.111758245266795e-06,
+      "loss": 0.824,
+      "step": 272
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.041789404643078e-06,
+      "loss": 0.8308,
+      "step": 273
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.972049596166707e-06,
+      "loss": 0.8239,
+      "step": 274
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.902542855160642e-06,
+      "loss": 0.798,
+      "step": 275
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.83327320346196e-06,
+      "loss": 0.8063,
+      "step": 276
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.76424464918915e-06,
+      "loss": 0.7963,
+      "step": 277
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.695461186510194e-06,
+      "loss": 0.8204,
+      "step": 278
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.626926795411447e-06,
+      "loss": 0.8204,
+      "step": 279
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.558645441467347e-06,
+      "loss": 0.7832,
+      "step": 280
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.490621075610958e-06,
+      "loss": 0.8548,
+      "step": 281
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.422857633905358e-06,
+      "loss": 0.8022,
+      "step": 282
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.355359037315893e-06,
+      "loss": 0.8975,
+      "step": 283
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.288129191483297e-06,
+      "loss": 0.8389,
+      "step": 284
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.221171986497686e-06,
+      "loss": 0.8377,
+      "step": 285
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.1544912966735e-06,
+      "loss": 0.7948,
+      "step": 286
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.0880909803252974e-06,
+      "loss": 0.7896,
+      "step": 287
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.021974879544522e-06,
+      "loss": 0.8582,
+      "step": 288
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.956146819977166e-06,
+      "loss": 0.842,
+      "step": 289
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.890610610602437e-06,
+      "loss": 0.8021,
+      "step": 290
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.82537004351234e-06,
+      "loss": 0.7894,
+      "step": 291
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.760428893692274e-06,
+      "loss": 0.8664,
+      "step": 292
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.695790918802577e-06,
+      "loss": 0.8016,
+      "step": 293
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.631459858961123e-06,
+      "loss": 0.8085,
+      "step": 294
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.567439436526897e-06,
+      "loss": 0.8079,
+      "step": 295
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.5037333558846155e-06,
+      "loss": 0.782,
+      "step": 296
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.440345303230377e-06,
+      "loss": 0.7886,
+      "step": 297
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.377278946358363e-06,
+      "loss": 0.8899,
+      "step": 298
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.314537934448628e-06,
+      "loss": 0.8182,
+      "step": 299
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.2521258978559324e-06,
+      "loss": 0.8336,
+      "step": 300
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.19004644789969e-06,
+      "loss": 0.8754,
+      "step": 301
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.128303176655002e-06,
+      "loss": 0.7593,
+      "step": 302
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.066899656744816e-06,
+      "loss": 0.8286,
+      "step": 303
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.005839441133198e-06,
+      "loss": 0.8598,
+      "step": 304
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.945126062919756e-06,
+      "loss": 0.8381,
+      "step": 305
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.884763035135205e-06,
+      "loss": 0.7859,
+      "step": 306
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.824753850538082e-06,
+      "loss": 0.8469,
+      "step": 307
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7651019814126656e-06,
+      "loss": 0.8343,
+      "step": 308
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7058108793680467e-06,
+      "loss": 0.8597,
+      "step": 309
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.646883975138421e-06,
+      "loss": 0.8567,
+      "step": 310
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.588324678384555e-06,
+      "loss": 0.7813,
+      "step": 311
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5301363774965256e-06,
+      "loss": 0.7459,
+      "step": 312
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.4723224393976353e-06,
+      "loss": 0.8251,
+      "step": 313
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.414886209349615e-06,
+      "loss": 0.8412,
+      "step": 314
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.357831010759026e-06,
+      "loss": 0.7613,
+      "step": 315
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.301160144984992e-06,
+      "loss": 0.8462,
+      "step": 316
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2448768911481577e-06,
+      "loss": 0.8351,
+      "step": 317
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1889845059409552e-06,
+      "loss": 0.8315,
+      "step": 318
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1334862234391627e-06,
+      "loss": 0.7783,
+      "step": 319
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.078385254914764e-06,
+      "loss": 0.7982,
+      "step": 320
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.023684788650154e-06,
+      "loss": 0.8098,
+      "step": 321
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9693879897536436e-06,
+      "loss": 0.8252,
+      "step": 322
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.91549799997632e-06,
+      "loss": 0.8619,
+      "step": 323
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8620179375302604e-06,
+      "loss": 0.8367,
+      "step": 324
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8089508969081e-06,
+      "loss": 0.7489,
+      "step": 325
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7562999487039822e-06,
+      "loss": 0.8152,
+      "step": 326
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7040681394358813e-06,
+      "loss": 0.8379,
+      "step": 327
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6522584913693295e-06,
+      "loss": 0.8188,
+      "step": 328
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6008740023425248e-06,
+      "loss": 0.796,
+      "step": 329
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5499176455928933e-06,
+      "loss": 0.8382,
+      "step": 330
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4993923695850307e-06,
+      "loss": 0.8543,
+      "step": 331
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.449301097840107e-06,
+      "loss": 0.8014,
+      "step": 332
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3996467287666914e-06,
+      "loss": 0.7993,
+      "step": 333
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3504321354930572e-06,
+      "loss": 0.7433,
+      "step": 334
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3016601657009364e-06,
+      "loss": 0.7676,
+      "step": 335
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.253333641460732e-06,
+      "loss": 0.784,
+      "step": 336
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.205455359068227e-06,
+      "loss": 0.7983,
+      "step": 337
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1580280888828e-06,
+      "loss": 0.8481,
+      "step": 338
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1110545751671074e-06,
+      "loss": 0.8255,
+      "step": 339
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0645375359283047e-06,
+      "loss": 0.78,
+      "step": 340
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0184796627607728e-06,
+      "loss": 0.7633,
+      "step": 341
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.972883620690366e-06,
+      "loss": 0.8385,
+      "step": 342
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9277520480202205e-06,
+      "loss": 0.7888,
+      "step": 343
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8830875561780904e-06,
+      "loss": 0.8004,
+      "step": 344
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8388927295652449e-06,
+      "loss": 0.8498,
+      "step": 345
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.795170125406921e-06,
+      "loss": 0.8315,
+      "step": 346
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7519222736043662e-06,
+      "loss": 0.8116,
+      "step": 347
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7091516765884464e-06,
+      "loss": 0.8113,
+      "step": 348
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6668608091748495e-06,
+      "loss": 0.817,
+      "step": 349
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.625052118420889e-06,
+      "loss": 0.7449,
+      "step": 350
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5837280234839013e-06,
+      "loss": 0.7696,
+      "step": 351
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.542890915481282e-06,
+      "loss": 0.8255,
+      "step": 352
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5025431573521209e-06,
+      "loss": 0.8188,
+      "step": 353
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4626870837204776e-06,
+      "loss": 0.8371,
+      "step": 354
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4233250007602873e-06,
+      "loss": 0.7561,
+      "step": 355
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3844591860619382e-06,
+      "loss": 0.7779,
+      "step": 356
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.346091888500466e-06,
+      "loss": 0.8593,
+      "step": 357
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.308225328105439e-06,
+      "loss": 0.7571,
+      "step": 358
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2708616959324893e-06,
+      "loss": 0.8681,
+      "step": 359
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2340031539365483e-06,
+      "loss": 0.7252,
+      "step": 360
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1976518348467426e-06,
+      "loss": 0.822,
+      "step": 361
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.161809842042988e-06,
+      "loss": 0.8157,
+      "step": 362
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1264792494342858e-06,
+      "loss": 0.8135,
+      "step": 363
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.091662101338714e-06,
+      "loss": 0.8323,
+      "step": 364
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0573604123651504e-06,
+      "loss": 0.7858,
+      "step": 365
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0235761672966903e-06,
+      "loss": 0.8419,
+      "step": 366
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.903113209758098e-07,
+      "loss": 0.7914,
+      "step": 367
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.575677981912457e-07,
+      "loss": 0.8128,
+      "step": 368
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.25347493566634e-07,
+      "loss": 0.78,
+      "step": 369
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.936522714508678e-07,
+      "loss": 0.813,
+      "step": 370
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.624839658102346e-07,
+      "loss": 0.8214,
+      "step": 371
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.318443801222919e-07,
+      "loss": 0.825,
+      "step": 372
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.017352872715078e-07,
+      "loss": 0.8359,
+      "step": 373
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.7215842944669e-07,
+      "loss": 0.7801,
+      "step": 374
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.431155180401705e-07,
+      "loss": 0.7953,
+      "step": 375
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.146082335487825e-07,
+      "loss": 0.7873,
+      "step": 376
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.866382254766158e-07,
+      "loss": 0.861,
+      "step": 377
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.592071122395849e-07,
+      "loss": 0.7912,
+      "step": 378
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.323164810717753e-07,
+      "loss": 0.7769,
+      "step": 379
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.059678879336006e-07,
+      "loss": 0.8244,
+      "step": 380
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.801628574217733e-07,
+      "loss": 0.8066,
+      "step": 381
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.549028826810888e-07,
+      "loss": 0.8034,
+      "step": 382
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.301894253180295e-07,
+      "loss": 0.7885,
+      "step": 383
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.060239153161872e-07,
+      "loss": 0.8412,
+      "step": 384
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.824077509535252e-07,
+      "loss": 0.7949,
+      "step": 385
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.5934229872146683e-07,
+      "loss": 0.7916,
+      "step": 386
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.3682889324583086e-07,
+      "loss": 0.7595,
+      "step": 387
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1486883720960436e-07,
+      "loss": 0.7135,
+      "step": 388
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.934634012775662e-07,
+      "loss": 0.795,
+      "step": 389
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.726138240227628e-07,
+      "loss": 0.8281,
+      "step": 390
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.5232131185484075e-07,
+      "loss": 0.8039,
+      "step": 391
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.325870389502439e-07,
+      "loss": 0.7719,
+      "step": 392
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1341214718426885e-07,
+      "loss": 0.8072,
+      "step": 393
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.9479774606499755e-07,
+      "loss": 0.8029,
+      "step": 394
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.7674491266909023e-07,
+      "loss": 0.7855,
+      "step": 395
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.5925469157947135e-07,
+      "loss": 0.8173,
+      "step": 396
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.423280948248841e-07,
+      "loss": 0.7435,
+      "step": 397
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.2596610182133328e-07,
+      "loss": 0.8452,
+      "step": 398
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1016965931541012e-07,
+      "loss": 0.7982,
+      "step": 399
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9493968132951456e-07,
+      "loss": 0.771,
+      "step": 400
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.802770491089667e-07,
+      "loss": 0.7866,
+      "step": 401
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.661826110710163e-07,
+      "loss": 0.8198,
+      "step": 402
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5265718275574658e-07,
+      "loss": 0.7537,
+      "step": 403
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3970154677889313e-07,
+      "loss": 0.7839,
+      "step": 404
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2731645278655448e-07,
+      "loss": 0.7849,
+      "step": 405
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1550261741181568e-07,
+      "loss": 0.8754,
+      "step": 406
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0426072423328382e-07,
+      "loss": 0.7954,
+      "step": 407
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.359142373553287e-08,
+      "loss": 0.8272,
+      "step": 408
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.349533327146719e-08,
+      "loss": 0.7997,
+      "step": 409
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.397303702659675e-08,
+      "loss": 0.7752,
+      "step": 410
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.50250859852375e-08,
+      "loss": 0.8586,
+      "step": 411
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.665199789862907e-08,
+      "loss": 0.8161,
+      "step": 412
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.88542572549755e-08,
+      "loss": 0.8268,
+      "step": 413
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.163231525141309e-08,
+      "loss": 0.8432,
+      "step": 414
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.498658976790248e-08,
+      "loss": 0.8082,
+      "step": 415
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.8917465343047956e-08,
+      "loss": 0.7384,
+      "step": 416
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.3425293151845274e-08,
+      "loss": 0.873,
+      "step": 417
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.851039098537122e-08,
+      "loss": 0.814,
+      "step": 418
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.4173043232380557e-08,
+      "loss": 0.8339,
+      "step": 419
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0413500862864745e-08,
+      "loss": 0.8343,
+      "step": 420
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.231981413520217e-09,
+      "loss": 0.733,
+      "step": 421
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.628668975166228e-09,
+      "loss": 0.817,
+      "step": 422
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.6037141820933752e-09,
+      "loss": 0.8674,
+      "step": 423
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.157234203341684e-09,
+      "loss": 0.8376,
+      "step": 424
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.8931273592824884e-10,
+      "loss": 0.803,
+      "step": 425
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0,
+      "loss": 0.8391,
+      "step": 426
+    },
+    {
+      "epoch": 1.0,
+      "step": 426,
+      "total_flos": 1.7489761962148495e+18,
+      "train_loss": 0.8527796469943624,
+      "train_runtime": 1012.9291,
+      "train_samples_per_second": 214.873,
+      "train_steps_per_second": 0.421
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 426,
+  "num_train_epochs": 1,
+  "save_steps": 5000,
+  "total_flos": 1.7489761962148495e+18,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca91a570a252195f778aed44093e543d34ea7877c46d110d835827391b873e90
+size 5307