dsakerkwq commited on
Commit
5bf83f0
1 Parent(s): e0770f5

End of training

Browse files
Files changed (2) hide show
  1. README.md +168 -0
  2. adapter_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,168 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ license: apache-2.0
4
+ base_model: unsloth/Qwen2.5-Coder-7B-Instruct
5
+ tags:
6
+ - axolotl
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: 9cd2803a-1298-4b5a-b9c4-dfca26236cbb
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
17
+ <details><summary>See axolotl config</summary>
18
+
19
+ axolotl version: `0.4.1`
20
+ ```yaml
21
+ adapter: lora
22
+ base_model: unsloth/Qwen2.5-Coder-7B-Instruct
23
+ bf16: auto
24
+ chat_template: llama3
25
+ cosine_min_lr_ratio: 0.1
26
+ data_processes: 4
27
+ dataset_prepared_path: null
28
+ datasets:
29
+ - data_files:
30
+ - 1745897a45c9b908_train_data.json
31
+ ds_type: json
32
+ field: content
33
+ num_proc: 4
34
+ path: /workspace/input_data/1745897a45c9b908_train_data.json
35
+ streaming: true
36
+ type: completion
37
+ debug: null
38
+ deepspeed: null
39
+ device_map: balanced
40
+ do_eval: true
41
+ early_stopping_patience: 1
42
+ eval_batch_size: 1
43
+ eval_sample_packing: false
44
+ eval_steps: 25
45
+ evaluation_strategy: steps
46
+ flash_attention: false
47
+ fp16: null
48
+ fsdp: null
49
+ fsdp_config: null
50
+ gradient_accumulation_steps: 16
51
+ gradient_checkpointing: true
52
+ group_by_length: true
53
+ hub_model_id: dsakerkwq/9cd2803a-1298-4b5a-b9c4-dfca26236cbb
54
+ hub_strategy: checkpoint
55
+ hub_token: null
56
+ learning_rate: 0.0001
57
+ load_in_4bit: false
58
+ load_in_8bit: false
59
+ local_rank: null
60
+ logging_steps: 1
61
+ lora_alpha: 64
62
+ lora_dropout: 0.05
63
+ lora_fan_in_fan_out: null
64
+ lora_model_dir: null
65
+ lora_r: 32
66
+ lora_target_linear: true
67
+ lora_target_modules:
68
+ - q_proj
69
+ - v_proj
70
+ lr_scheduler: cosine
71
+ max_grad_norm: 1.0
72
+ max_memory:
73
+ 0: 75GB
74
+ 1: 75GB
75
+ 2: 75GB
76
+ 3: 75GB
77
+ max_steps: 50
78
+ micro_batch_size: 2
79
+ mixed_precision: bf16
80
+ mlflow_experiment_name: /tmp/1745897a45c9b908_train_data.json
81
+ model_type: AutoModelForCausalLM
82
+ num_epochs: 3
83
+ optim_args:
84
+ adam_beta1: 0.9
85
+ adam_beta2: 0.95
86
+ adam_epsilon: 1e-5
87
+ optimizer: adamw_torch
88
+ output_dir: miner_id_24
89
+ pad_to_sequence_len: true
90
+ resume_from_checkpoint: null
91
+ s2_attention: null
92
+ sample_packing: false
93
+ save_steps: 25
94
+ save_strategy: steps
95
+ sequence_len: 2048
96
+ strict: false
97
+ tf32: false
98
+ tokenizer_type: AutoTokenizer
99
+ torch_compile: false
100
+ train_on_inputs: false
101
+ trust_remote_code: true
102
+ val_set_size: 50
103
+ wandb_entity: null
104
+ wandb_mode: online
105
+ wandb_name: 9cd2803a-1298-4b5a-b9c4-dfca26236cbb
106
+ wandb_project: Public_TuningSN
107
+ wandb_runid: 9cd2803a-1298-4b5a-b9c4-dfca26236cbb
108
+ warmup_ratio: 0.04
109
+ weight_decay: 0.01
110
+ xformers_attention: null
111
+
112
+ ```
113
+
114
+ </details><br>
115
+
116
+ # 9cd2803a-1298-4b5a-b9c4-dfca26236cbb
117
+
118
+ This model is a fine-tuned version of [unsloth/Qwen2.5-Coder-7B-Instruct](https://huggingface.co/unsloth/Qwen2.5-Coder-7B-Instruct) on the None dataset.
119
+ It achieves the following results on the evaluation set:
120
+ - Loss: nan
121
+
122
+ ## Model description
123
+
124
+ More information needed
125
+
126
+ ## Intended uses & limitations
127
+
128
+ More information needed
129
+
130
+ ## Training and evaluation data
131
+
132
+ More information needed
133
+
134
+ ## Training procedure
135
+
136
+ ### Training hyperparameters
137
+
138
+ The following hyperparameters were used during training:
139
+ - learning_rate: 0.0001
140
+ - train_batch_size: 2
141
+ - eval_batch_size: 1
142
+ - seed: 42
143
+ - distributed_type: multi-GPU
144
+ - num_devices: 4
145
+ - gradient_accumulation_steps: 16
146
+ - total_train_batch_size: 128
147
+ - total_eval_batch_size: 4
148
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
149
+ - lr_scheduler_type: cosine
150
+ - lr_scheduler_warmup_steps: 2
151
+ - training_steps: 50
152
+
153
+ ### Training results
154
+
155
+ | Training Loss | Epoch | Step | Validation Loss |
156
+ |:-------------:|:------:|:----:|:---------------:|
157
+ | 0.4794 | 0.0044 | 1 | nan |
158
+ | 0.0 | 0.1098 | 25 | nan |
159
+ | 0.0 | 0.2197 | 50 | nan |
160
+
161
+
162
+ ### Framework versions
163
+
164
+ - PEFT 0.13.2
165
+ - Transformers 4.46.0
166
+ - Pytorch 2.5.0+cu124
167
+ - Datasets 3.0.1
168
+ - Tokenizers 0.20.1
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:81223561dd8650cc0df15b59a89aa280faa4b3baa8beb0fd62378a4d54145fe6
3
+ size 323103018