Model save
Browse files- README.md +143 -0
- generation_config.json +6 -0
- model.safetensors +1 -1
README.md
ADDED
@@ -0,0 +1,143 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: other
|
3 |
+
base_model: deepseek-ai/deepseek-coder-1.3b-base
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
model-index:
|
7 |
+
- name: deepseek_coder_1.3b_typescript
|
8 |
+
results: []
|
9 |
+
---
|
10 |
+
|
11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
12 |
+
should probably proofread and complete it, then remove this comment. -->
|
13 |
+
|
14 |
+
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
15 |
+
<details><summary>See axolotl config</summary>
|
16 |
+
|
17 |
+
axolotl version: `0.3.0`
|
18 |
+
```yaml
|
19 |
+
base_model: deepseek-ai/deepseek-coder-1.3b-base
|
20 |
+
model_type: AutoModelForCausalLM
|
21 |
+
trust_remote_code: true
|
22 |
+
load_in_8bit: false
|
23 |
+
load_in_4bit: false
|
24 |
+
strict: false
|
25 |
+
|
26 |
+
|
27 |
+
datasets:
|
28 |
+
- path: CodeGPTPlus/typescript-0-500000-seq1024
|
29 |
+
type: completion
|
30 |
+
field: text
|
31 |
+
#dataset_prepared_path:
|
32 |
+
|
33 |
+
#pretraining_dataset: CodeGPTPlus/typescript-0-500000-seq1024
|
34 |
+
|
35 |
+
val_set_size: 0.001
|
36 |
+
output_dir: ./fft-out
|
37 |
+
|
38 |
+
sequence_len: 1024
|
39 |
+
|
40 |
+
adapter:
|
41 |
+
lora_model_dir:
|
42 |
+
lora_r:
|
43 |
+
lora_alpha:
|
44 |
+
lora_dropout:
|
45 |
+
lora_target_linear:
|
46 |
+
lora_fan_in_fan_out:
|
47 |
+
lora_modules_to_save:
|
48 |
+
|
49 |
+
wandb_project: deepseek_1.3_fft
|
50 |
+
wandb_entity:
|
51 |
+
wandb_watch:
|
52 |
+
wandb_name: aws_a10g
|
53 |
+
wandb_log_model: end
|
54 |
+
|
55 |
+
|
56 |
+
gradient_accumulation_steps: 2
|
57 |
+
micro_batch_size: 20
|
58 |
+
num_epochs: 1
|
59 |
+
#max_steps: 1 # REMOVE IT
|
60 |
+
optimizer: adamw_bnb_8bit
|
61 |
+
adam_beta1: 0.9
|
62 |
+
adam_beta2: 0.999
|
63 |
+
adam_epsilon: 0.000001
|
64 |
+
max_grad_norm: 1.0
|
65 |
+
weight_decay: 0.1
|
66 |
+
lr_scheduler: cosine
|
67 |
+
learning_rate: 0.00002
|
68 |
+
train_on_inputs: false
|
69 |
+
group_by_length: false
|
70 |
+
bf16: true
|
71 |
+
fp16: false
|
72 |
+
tf32: false
|
73 |
+
gradient_checkpointing: true
|
74 |
+
early_stopping_patience:
|
75 |
+
resume_from_checkpoint:
|
76 |
+
local_rank:
|
77 |
+
logging_steps: 1
|
78 |
+
xformers_attention:
|
79 |
+
flash_attention: true
|
80 |
+
|
81 |
+
loss_watchdog_threshold: 5.0
|
82 |
+
loss_watchdog_patience: 3
|
83 |
+
|
84 |
+
hub_model_id: CodeGPTPlus/deepseek_coder_1.3b_typescript
|
85 |
+
hub_strategy: every_save
|
86 |
+
warmup_ratio: 0.01
|
87 |
+
evals_per_epoch: 20
|
88 |
+
saves_per_epoch: 3
|
89 |
+
debug:
|
90 |
+
deepspeed:
|
91 |
+
|
92 |
+
fsdp:
|
93 |
+
fsdp_config:
|
94 |
+
special_tokens:
|
95 |
+
bos_token: "<|begin▁of▁sentence|>"
|
96 |
+
eos_token: "<|end▁of▁sentence|>"
|
97 |
+
pad_token: "<|end▁of▁sentence|>"
|
98 |
+
# fim_prefix: "<|fim▁begin|>"
|
99 |
+
# fim_middle: "<|fim▁hole|>"
|
100 |
+
# fim_suffix: "<|fim▁end|>"
|
101 |
+
|
102 |
+
```
|
103 |
+
|
104 |
+
</details><br>
|
105 |
+
|
106 |
+
# deepseek_coder_1.3b_typescript
|
107 |
+
|
108 |
+
This model is a fine-tuned version of [deepseek-ai/deepseek-coder-1.3b-base](https://huggingface.co/deepseek-ai/deepseek-coder-1.3b-base) on an unknown dataset.
|
109 |
+
|
110 |
+
## Model description
|
111 |
+
|
112 |
+
More information needed
|
113 |
+
|
114 |
+
## Intended uses & limitations
|
115 |
+
|
116 |
+
More information needed
|
117 |
+
|
118 |
+
## Training and evaluation data
|
119 |
+
|
120 |
+
More information needed
|
121 |
+
|
122 |
+
## Training procedure
|
123 |
+
|
124 |
+
### Training hyperparameters
|
125 |
+
|
126 |
+
The following hyperparameters were used during training:
|
127 |
+
- learning_rate: 2e-05
|
128 |
+
- train_batch_size: 20
|
129 |
+
- eval_batch_size: 20
|
130 |
+
- seed: 42
|
131 |
+
- gradient_accumulation_steps: 2
|
132 |
+
- total_train_batch_size: 40
|
133 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
|
134 |
+
- lr_scheduler_type: cosine
|
135 |
+
- lr_scheduler_warmup_steps: 261
|
136 |
+
- num_epochs: 1
|
137 |
+
|
138 |
+
### Framework versions
|
139 |
+
|
140 |
+
- Transformers 4.37.0.dev0
|
141 |
+
- Pytorch 2.0.1+cu118
|
142 |
+
- Datasets 2.16.1
|
143 |
+
- Tokenizers 0.15.0
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 32013,
|
4 |
+
"eos_token_id": 32014,
|
5 |
+
"transformers_version": "4.37.0.dev0"
|
6 |
+
}
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3234034400
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fee5059bda6faea9f81360b5b876ec081744de1940b536f62c1b6945a9876d2c
|
3 |
size 3234034400
|