Upload folder using huggingface_hub

README.md

@@ -0,0 +1,59 @@
+---
+base_model: Qwen/Qwen-1_8B-Chat
+library_name: peft
+license: other
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+model-index:
+- name: train_2024-07-16-18-56-15
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# train_2024-07-16-18-56-15
+
+This model is a fine-tuned version of [Qwen/Qwen-1_8B-Chat](https://huggingface.co/Qwen/Qwen-1_8B-Chat) on the glaive_toolcall_en_demo dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 5e-05
+- train_batch_size: 2
+- eval_batch_size: 8
+- seed: 42
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- num_epochs: 3.0
+- mixed_precision_training: Native AMP
+
+### Training results
+
+
+
+### Framework versions
+
+- PEFT 0.11.1
+- Transformers 4.41.2
+- Pytorch 2.3.0+cu121
+- Datasets 2.20.0
+- Tokenizers 0.19.1

adapter_config.json

@@ -0,0 +1,31 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen-1_8B-Chat",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "c_attn",
+    "c_proj",
+    "w1",
+    "w2"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afe14312e77d17e1f123d737ac5109b3156574658b1cecd671446dce7b9722b5
+size 26867400

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 2.88,
+    "num_input_tokens_seen": 530864,
+    "total_flos": 4880889936150528.0,
+    "train_loss": 0.5967088187182391,
+    "train_runtime": 437.0311,
+    "train_samples_per_second": 2.059,
+    "train_steps_per_second": 0.124
+}

llamaboard_config.yaml

@@ -0,0 +1,65 @@
+top.booster: auto
+top.checkpoint_path: []
+top.finetuning_type: lora
+top.model_name: Qwen-1.8B-Chat
+top.quantization_bit: none
+top.quantization_method: bitsandbytes
+top.rope_scaling: none
+top.template: qwen
+top.visual_inputs: false
+train.additional_target: ''
+train.badam_mode: layer
+train.badam_switch_interval: 50
+train.badam_switch_mode: ascending
+train.badam_update_ratio: 0.05
+train.batch_size: 2
+train.compute_type: fp16
+train.create_new_adapter: false
+train.cutoff_len: 1024
+train.dataset:
+- glaive_toolcall_en_demo
+train.dataset_dir: data
+train.ds_offload: false
+train.ds_stage: none
+train.freeze_extra_modules: ''
+train.freeze_trainable_layers: 2
+train.freeze_trainable_modules: all
+train.galore_rank: 16
+train.galore_scale: 0.25
+train.galore_target: all
+train.galore_update_interval: 200
+train.gradient_accumulation_steps: 8
+train.learning_rate: 5e-5
+train.logging_steps: 5
+train.lora_alpha: 16
+train.lora_dropout: 0
+train.lora_rank: 8
+train.lora_target: ''
+train.loraplus_lr_ratio: 0
+train.lr_scheduler_type: cosine
+train.max_grad_norm: '1.0'
+train.max_samples: '100000'
+train.neat_packing: false
+train.neftune_alpha: 0
+train.num_train_epochs: '3.0'
+train.optim: adamw_torch
+train.packing: false
+train.ppo_score_norm: false
+train.ppo_whiten_rewards: false
+train.pref_beta: 0.1
+train.pref_ftx: 0
+train.pref_loss: sigmoid
+train.report_to: false
+train.resize_vocab: false
+train.reward_model: null
+train.save_steps: 100
+train.shift_attn: false
+train.training_stage: Supervised Fine-Tuning
+train.use_badam: false
+train.use_dora: false
+train.use_galore: false
+train.use_llama_pro: false
+train.use_pissa: false
+train.use_rslora: false
+train.val_size: 0
+train.warmup_steps: 0

running_log.txt

@@ -0,0 +1,208 @@
+[INFO|parser.py:344] 2024-07-16 19:25:29,726 >> Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: False, compute dtype: torch.float16
+
+[INFO|tokenization_utils_base.py:2108] 2024-07-16 19:25:32,413 >> loading file qwen.tiktoken from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/qwen.tiktoken
+
+[INFO|tokenization_utils_base.py:2108] 2024-07-16 19:25:32,413 >> loading file added_tokens.json from cache at None
+
+[INFO|tokenization_utils_base.py:2108] 2024-07-16 19:25:32,413 >> loading file special_tokens_map.json from cache at None
+
+[INFO|tokenization_utils_base.py:2108] 2024-07-16 19:25:32,413 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/tokenizer_config.json
+
+[INFO|tokenization_utils_base.py:2108] 2024-07-16 19:25:32,414 >> loading file tokenizer.json from cache at None
+
+[INFO|template.py:268] 2024-07-16 19:25:32,773 >> Add eos token: <|im_end|>
+
+[INFO|template.py:372] 2024-07-16 19:25:32,773 >> Add pad token: <|im_end|>
+
+[INFO|loader.py:52] 2024-07-16 19:25:32,774 >> Loading dataset glaive_toolcall_en_demo.json...
+
+[INFO|configuration_utils.py:733] 2024-07-16 19:26:23,755 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/config.json
+
+[INFO|configuration_utils.py:733] 2024-07-16 19:26:24,494 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/config.json
+
+[INFO|configuration_utils.py:796] 2024-07-16 19:26:24,495 >> Model config QWenConfig {
+  "_name_or_path": "Qwen/Qwen-1_8B-Chat",
+  "architectures": [
+    "QWenLMHeadModel"
+  ],
+  "attn_dropout_prob": 0.0,
+  "auto_map": {
+    "AutoConfig": "Qwen/Qwen-1_8B-Chat--configuration_qwen.QWenConfig",
+    "AutoModelForCausalLM": "Qwen/Qwen-1_8B-Chat--modeling_qwen.QWenLMHeadModel"
+  },
+  "bf16": false,
+  "emb_dropout_prob": 0.0,
+  "fp16": false,
+  "fp32": false,
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "kv_channels": 128,
+  "layer_norm_epsilon": 1e-06,
+  "max_position_embeddings": 8192,
+  "model_type": "qwen",
+  "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,
+  "softmax_in_fp32": false,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "QWenTokenizer",
+  "transformers_version": "4.41.2",
+  "use_cache": true,
+  "use_cache_kernel": false,
+  "use_cache_quantization": false,
+  "use_dynamic_ntk": true,
+  "use_flash_attn": "auto",
+  "use_logn_attn": true,
+  "vocab_size": 151936
+}
+
+
+[INFO|modeling_utils.py:3474] 2024-07-16 19:26:26,974 >> loading weights file model.safetensors from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/model.safetensors.index.json
+
+[INFO|modeling_utils.py:1519] 2024-07-16 19:26:45,032 >> Instantiating QWenLMHeadModel model under default dtype torch.float16.
+
+[INFO|configuration_utils.py:962] 2024-07-16 19:26:45,034 >> Generate config GenerationConfig {}
+
+
+[INFO|modeling_utils.py:4280] 2024-07-16 19:26:51,937 >> All model checkpoint weights were used when initializing QWenLMHeadModel.
+
+
+[INFO|modeling_utils.py:4288] 2024-07-16 19:26:51,937 >> All the weights of QWenLMHeadModel were initialized from the model checkpoint at Qwen/Qwen-1_8B-Chat.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use QWenLMHeadModel for predictions without further training.
+
+[INFO|configuration_utils.py:917] 2024-07-16 19:26:52,423 >> loading configuration file generation_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/generation_config.json
+
+[INFO|configuration_utils.py:962] 2024-07-16 19:26:52,424 >> Generate config GenerationConfig {
+  "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
+}
+
+
+[WARNING|checkpointing.py:70] 2024-07-16 19:26:52,440 >> You are using the old GC format, some features (e.g. BAdam) will be invalid.
+
+[INFO|checkpointing.py:103] 2024-07-16 19:26:52,440 >> Gradient checkpointing enabled.
+
+[INFO|attention.py:86] 2024-07-16 19:26:52,440 >> Using vanilla attention implementation.
+
+[INFO|adapter.py:302] 2024-07-16 19:26:52,441 >> Upcasting trainable params to float32.
+
+[INFO|adapter.py:158] 2024-07-16 19:26:52,441 >> Fine-tuning method: LoRA
+
+[INFO|misc.py:51] 2024-07-16 19:26:52,442 >> Found linear modules: c_attn,c_proj,w1,w2
+
+[INFO|loader.py:196] 2024-07-16 19:26:53,145 >> trainable params: 6,709,248 || all params: 1,843,537,920 || trainable%: 0.3639
+
+[INFO|trainer.py:641] 2024-07-16 19:26:53,161 >> Using auto half precision backend
+
+[INFO|trainer.py:2078] 2024-07-16 19:26:54,481 >> ***** Running training *****
+
+[INFO|trainer.py:2079] 2024-07-16 19:26:54,481 >>   Num examples = 300
+
+[INFO|trainer.py:2080] 2024-07-16 19:26:54,481 >>   Num Epochs = 3
+
+[INFO|trainer.py:2081] 2024-07-16 19:26:54,481 >>   Instantaneous batch size per device = 2
+
+[INFO|trainer.py:2084] 2024-07-16 19:26:54,481 >>   Total train batch size (w. parallel, distributed & accumulation) = 16
+
+[INFO|trainer.py:2085] 2024-07-16 19:26:54,481 >>   Gradient Accumulation steps = 8
+
+[INFO|trainer.py:2086] 2024-07-16 19:26:54,482 >>   Total optimization steps = 54
+
+[INFO|trainer.py:2087] 2024-07-16 19:26:54,484 >>   Number of trainable parameters = 6,709,248
+
+[INFO|callbacks.py:310] 2024-07-16 19:27:37,133 >> {'loss': 0.6756, 'learning_rate': 4.8950e-05, 'epoch': 0.27, 'throughput': 1193.09}
+
+[INFO|callbacks.py:310] 2024-07-16 19:28:17,112 >> {'loss': 0.6799, 'learning_rate': 4.5887e-05, 'epoch': 0.53, 'throughput': 1200.41}
+
+[INFO|callbacks.py:310] 2024-07-16 19:28:56,484 >> {'loss': 0.6995, 'learning_rate': 4.1070e-05, 'epoch': 0.80, 'throughput': 1208.30}
+
+[INFO|callbacks.py:310] 2024-07-16 19:29:36,117 >> {'loss': 0.6313, 'learning_rate': 3.4902e-05, 'epoch': 1.07, 'throughput': 1209.98}
+
+[INFO|callbacks.py:310] 2024-07-16 19:30:17,804 >> {'loss': 0.5683, 'learning_rate': 2.7902e-05, 'epoch': 1.33, 'throughput': 1211.98}
+
+[INFO|callbacks.py:310] 2024-07-16 19:30:57,971 >> {'loss': 0.4988, 'learning_rate': 2.0659e-05, 'epoch': 1.60, 'throughput': 1214.64}
+
+[INFO|callbacks.py:310] 2024-07-16 19:31:38,903 >> {'loss': 0.5748, 'learning_rate': 1.3780e-05, 'epoch': 1.87, 'throughput': 1215.63}
+
+[INFO|callbacks.py:310] 2024-07-16 19:32:15,869 >> {'loss': 0.5793, 'learning_rate': 7.8440e-06, 'epoch': 2.13, 'throughput': 1214.06}
+
+[INFO|callbacks.py:310] 2024-07-16 19:32:55,941 >> {'loss': 0.5500, 'learning_rate': 3.3494e-06, 'epoch': 2.40, 'throughput': 1214.25}
+
+[INFO|callbacks.py:310] 2024-07-16 19:33:38,528 >> {'loss': 0.5715, 'learning_rate': 6.7388e-07, 'epoch': 2.67, 'throughput': 1214.09}
+
+[INFO|trainer.py:2329] 2024-07-16 19:34:11,515 >> 
+
+Training completed. Do not forget to share your model on huggingface.co/models =)
+
+
+
+[INFO|trainer.py:3410] 2024-07-16 19:34:11,517 >> Saving model checkpoint to saves/Qwen-1.8B-Chat/lora/train_2024-07-16-18-56-15
+
+[INFO|configuration_utils.py:733] 2024-07-16 19:34:12,032 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen-1_8B-Chat/snapshots/1d0f68de57b88cfde81f3c3e537f24464d889081/config.json
+
+[INFO|configuration_utils.py:796] 2024-07-16 19:34:12,033 >> Model config QWenConfig {
+  "architectures": [
+    "QWenLMHeadModel"
+  ],
+  "attn_dropout_prob": 0.0,
+  "auto_map": {
+    "AutoConfig": "Qwen/Qwen-1_8B-Chat--configuration_qwen.QWenConfig",
+    "AutoModelForCausalLM": "Qwen/Qwen-1_8B-Chat--modeling_qwen.QWenLMHeadModel"
+  },
+  "bf16": false,
+  "emb_dropout_prob": 0.0,
+  "fp16": false,
+  "fp32": false,
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "kv_channels": 128,
+  "layer_norm_epsilon": 1e-06,
+  "max_position_embeddings": 8192,
+  "model_type": "qwen",
+  "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,
+  "softmax_in_fp32": false,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "QWenTokenizer",
+  "transformers_version": "4.41.2",
+  "use_cache": true,
+  "use_cache_kernel": false,
+  "use_cache_quantization": false,
+  "use_dynamic_ntk": true,
+  "use_flash_attn": "auto",
+  "use_logn_attn": true,
+  "vocab_size": 151936
+}
+
+
+[INFO|tokenization_utils_base.py:2513] 2024-07-16 19:34:12,174 >> tokenizer config file saved in saves/Qwen-1.8B-Chat/lora/train_2024-07-16-18-56-15/tokenizer_config.json
+
+[INFO|tokenization_utils_base.py:2522] 2024-07-16 19:34:12,174 >> Special tokens file saved in saves/Qwen-1.8B-Chat/lora/train_2024-07-16-18-56-15/special_tokens_map.json
+
+[WARNING|ploting.py:89] 2024-07-16 19:34:12,511 >> No metric eval_loss to plot.
+
+[WARNING|ploting.py:89] 2024-07-16 19:34:12,512 >> No metric eval_accuracy to plot.
+
+[INFO|modelcard.py:450] 2024-07-16 19:34:12,513 >> Dropping the following result as it does not have all the necessary fields:
+{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}
+

special_tokens_map.json

@@ -0,0 +1,10 @@
+{
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|im_end|>"
+}

tokenization_qwen.py

@@ -0,0 +1,276 @@
+# 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)))
+# changed to use actual index to avoid misconfiguration with vocabulary expansion
+SPECIAL_START_ID = 151643
+SPECIAL_TOKENS = tuple(
+    enumerate(
+        (
+            (
+                ENDOFTEXT,
+                IMSTART,
+                IMEND,
+            )
+            + EXTRAS
+        ),
+        start=SPECIAL_START_ID,
+    )
+)
+SPECIAL_TOKENS_SET = set(t for i, t in SPECIAL_TOKENS)
+
+
+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",
+        extra_vocab_file=None,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        # how to handle errors in decoding UTF-8 byte sequences
+        # use ignore if you are in streaming inference
+        self.errors = errors  
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: Dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in SPECIAL_TOKENS
+        }
+
+        # try load extra vocab from file
+        if extra_vocab_file is not None:
+            used_ids = set(self.mergeable_ranks.values()) | set(self.special_tokens.values())
+            extra_mergeable_ranks = _load_tiktoken_bpe(extra_vocab_file)
+            for token, index in extra_mergeable_ranks.items():
+                if token in self.mergeable_ranks:
+                    logger.info(f"extra token {token} exists, skipping")
+                    continue
+                if index in used_ids:
+                    logger.info(f'the index {index} for extra token {token} exists, skipping')
+                    continue
+                self.mergeable_ranks[token] = index
+            # the index may be sparse after this, but don't worry tiktoken.Encoding will handle this
+
+        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 __getstate__(self):
+        # for pickle lovers
+        state = self.__dict__.copy()
+        del state["tokenizer"]
+        return state
+
+    def __setstate__(self, state):
+        # tokenizer is not python native; don't pass it; rebuild it
+        self.__dict__.update(state)
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        self.tokenizer = enc
+
+    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_SET:
+                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 +1,17 @@
+{
+  "added_tokens_decoder": {},
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+    ]
+  },
+  "chat_template": "{% set system_message = 'You are a helpful assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<|im_start|>system\n' + system_message + '<|im_end|>\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|im_start|>user\n' + content + '<|im_end|>\n<|im_start|>assistant\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|im_end|>' + '\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|im_end|>",
+  "model_max_length": 8192,
+  "pad_token": "<|im_end|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "QWenTokenizer"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 2.88,
+    "num_input_tokens_seen": 530864,
+    "total_flos": 4880889936150528.0,
+    "train_loss": 0.5967088187182391,
+    "train_runtime": 437.0311,
+    "train_samples_per_second": 2.059,
+    "train_steps_per_second": 0.124
+}

trainer_log.jsonl

@@ -0,0 +1,11 @@
+{"current_steps": 5, "total_steps": 54, "loss": 0.6756, "learning_rate": 4.894973780788722e-05, "epoch": 0.26666666666666666, "percentage": 9.26, "elapsed_time": "0:00:42", "remaining_time": "0:06:57", "throughput": "1193.09", "total_tokens": 50880}
+{"current_steps": 10, "total_steps": 54, "loss": 0.6799, "learning_rate": 4.588719528532342e-05, "epoch": 0.5333333333333333, "percentage": 18.52, "elapsed_time": "0:01:22", "remaining_time": "0:06:03", "throughput": "1200.41", "total_tokens": 99184}
+{"current_steps": 15, "total_steps": 54, "loss": 0.6995, "learning_rate": 4.1069690242163484e-05, "epoch": 0.8, "percentage": 27.78, "elapsed_time": "0:02:01", "remaining_time": "0:05:17", "throughput": "1208.30", "total_tokens": 147408}
+{"current_steps": 20, "total_steps": 54, "loss": 0.6313, "learning_rate": 3.490199415097892e-05, "epoch": 1.0666666666666667, "percentage": 37.04, "elapsed_time": "0:02:41", "remaining_time": "0:04:34", "throughput": "1209.98", "total_tokens": 195568}
+{"current_steps": 25, "total_steps": 54, "loss": 0.5683, "learning_rate": 2.7902322853130757e-05, "epoch": 1.3333333333333333, "percentage": 46.3, "elapsed_time": "0:03:23", "remaining_time": "0:03:55", "throughput": "1211.98", "total_tokens": 246416}
+{"current_steps": 30, "total_steps": 54, "loss": 0.4988, "learning_rate": 2.0658795558326743e-05, "epoch": 1.6, "percentage": 55.56, "elapsed_time": "0:04:03", "remaining_time": "0:03:14", "throughput": "1214.64", "total_tokens": 295744}
+{"current_steps": 35, "total_steps": 54, "loss": 0.5748, "learning_rate": 1.3780020494988446e-05, "epoch": 1.8666666666666667, "percentage": 64.81, "elapsed_time": "0:04:44", "remaining_time": "0:02:34", "throughput": "1215.63", "total_tokens": 345744}
+{"current_steps": 40, "total_steps": 54, "loss": 0.5793, "learning_rate": 7.843959053281663e-06, "epoch": 2.1333333333333333, "percentage": 74.07, "elapsed_time": "0:05:21", "remaining_time": "0:01:52", "throughput": "1214.06", "total_tokens": 390176}
+{"current_steps": 45, "total_steps": 54, "loss": 0.55, "learning_rate": 3.3493649053890326e-06, "epoch": 2.4, "percentage": 83.33, "elapsed_time": "0:06:01", "remaining_time": "0:01:12", "throughput": "1214.25", "total_tokens": 438896}
+{"current_steps": 50, "total_steps": 54, "loss": 0.5715, "learning_rate": 6.738782355044049e-07, "epoch": 2.6666666666666665, "percentage": 92.59, "elapsed_time": "0:06:44", "remaining_time": "0:00:32", "throughput": "1214.09", "total_tokens": 490544}
+{"current_steps": 54, "total_steps": 54, "epoch": 2.88, "percentage": 100.0, "elapsed_time": "0:07:17", "remaining_time": "0:00:00", "throughput": "1214.71", "total_tokens": 530864}

trainer_state.json

@@ -0,0 +1,123 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.88,
+  "eval_steps": 500,
+  "global_step": 54,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.26666666666666666,
+      "grad_norm": 0.4435124099254608,
+      "learning_rate": 4.894973780788722e-05,
+      "loss": 0.6756,
+      "num_input_tokens_seen": 50880,
+      "step": 5
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 0.5012878775596619,
+      "learning_rate": 4.588719528532342e-05,
+      "loss": 0.6799,
+      "num_input_tokens_seen": 99184,
+      "step": 10
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.31872692704200745,
+      "learning_rate": 4.1069690242163484e-05,
+      "loss": 0.6995,
+      "num_input_tokens_seen": 147408,
+      "step": 15
+    },
+    {
+      "epoch": 1.0666666666666667,
+      "grad_norm": 0.3336561322212219,
+      "learning_rate": 3.490199415097892e-05,
+      "loss": 0.6313,
+      "num_input_tokens_seen": 195568,
+      "step": 20
+    },
+    {
+      "epoch": 1.3333333333333333,
+      "grad_norm": 0.24536238610744476,
+      "learning_rate": 2.7902322853130757e-05,
+      "loss": 0.5683,
+      "num_input_tokens_seen": 246416,
+      "step": 25
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 0.2896718978881836,
+      "learning_rate": 2.0658795558326743e-05,
+      "loss": 0.4988,
+      "num_input_tokens_seen": 295744,
+      "step": 30
+    },
+    {
+      "epoch": 1.8666666666666667,
+      "grad_norm": 0.2828458249568939,
+      "learning_rate": 1.3780020494988446e-05,
+      "loss": 0.5748,
+      "num_input_tokens_seen": 345744,
+      "step": 35
+    },
+    {
+      "epoch": 2.1333333333333333,
+      "grad_norm": 0.37326768040657043,
+      "learning_rate": 7.843959053281663e-06,
+      "loss": 0.5793,
+      "num_input_tokens_seen": 390176,
+      "step": 40
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 0.28728172183036804,
+      "learning_rate": 3.3493649053890326e-06,
+      "loss": 0.55,
+      "num_input_tokens_seen": 438896,
+      "step": 45
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "grad_norm": 0.275715172290802,
+      "learning_rate": 6.738782355044049e-07,
+      "loss": 0.5715,
+      "num_input_tokens_seen": 490544,
+      "step": 50
+    },
+    {
+      "epoch": 2.88,
+      "num_input_tokens_seen": 530864,
+      "step": 54,
+      "total_flos": 4880889936150528.0,
+      "train_loss": 0.5967088187182391,
+      "train_runtime": 437.0311,
+      "train_samples_per_second": 2.059,
+      "train_steps_per_second": 0.124
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 54,
+  "num_input_tokens_seen": 530864,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4880889936150528.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:653a7a5eb95c010e25e2d17462f48539df2aa0b66714044c6de1173e1f47d33d
+size 5304

training_args.yaml

@@ -0,0 +1,33 @@
+cutoff_len: 1024
+dataset: glaive_toolcall_en_demo
+dataset_dir: data
+ddp_timeout: 180000000
+do_train: true
+finetuning_type: lora
+flash_attn: auto
+fp16: true
+gradient_accumulation_steps: 8
+include_num_input_tokens_seen: true
+learning_rate: 5.0e-05
+logging_steps: 5
+lora_alpha: 16
+lora_dropout: 0
+lora_rank: 8
+lora_target: all
+lr_scheduler_type: cosine
+max_grad_norm: 1.0
+max_samples: 100000
+model_name_or_path: Qwen/Qwen-1_8B-Chat
+num_train_epochs: 3.0
+optim: adamw_torch
+output_dir: saves/Qwen-1.8B-Chat/lora/train_2024-07-16-18-56-15
+packing: false
+per_device_train_batch_size: 2
+plot_loss: true
+preprocessing_num_workers: 16
+quantization_method: bitsandbytes
+report_to: none
+save_steps: 100
+stage: sft
+template: qwen
+warmup_steps: 0