End of training
Browse files- README.md +29 -10
- adapter_model.bin +2 -2
README.md
CHANGED
@@ -2,10 +2,11 @@
|
|
2 |
license: other
|
3 |
library_name: peft
|
4 |
tags:
|
|
|
5 |
- generated_from_trainer
|
6 |
base_model: deepseek-ai/deepseek-coder-6.7b-instruct
|
7 |
model-index:
|
8 |
-
- name:
|
9 |
results: []
|
10 |
---
|
11 |
|
@@ -27,10 +28,16 @@ strict: false
|
|
27 |
|
28 |
datasets:
|
29 |
- path: vdaita/editpackft_inst_code
|
|
|
30 |
type: oasst
|
31 |
dataset_prepared_path:
|
32 |
-
|
33 |
-
|
|
|
|
|
|
|
|
|
|
|
34 |
|
35 |
sequence_len: 4096
|
36 |
sample_packing: true
|
@@ -46,8 +53,15 @@ lora_dropout: 0.05
|
|
46 |
lora_target_linear: true
|
47 |
lora_fan_in_fan_out:
|
48 |
|
|
|
|
|
|
|
|
|
49 |
wandb_project: huggingface
|
50 |
-
wandb_log_model: axolotl-dscoder-code-
|
|
|
|
|
|
|
51 |
|
52 |
gradient_accumulation_steps: 4
|
53 |
micro_batch_size: 2
|
@@ -80,15 +94,20 @@ weight_decay: 0.0
|
|
80 |
fsdp:
|
81 |
fsdp_config:
|
82 |
|
|
|
|
|
|
|
|
|
|
|
83 |
```
|
84 |
|
85 |
</details><br>
|
86 |
|
87 |
-
#
|
88 |
|
89 |
This model is a fine-tuned version of [deepseek-ai/deepseek-coder-6.7b-instruct](https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-instruct) on the None dataset.
|
90 |
It achieves the following results on the evaluation set:
|
91 |
-
- Loss: 0.
|
92 |
|
93 |
## Model description
|
94 |
|
@@ -125,10 +144,10 @@ The following hyperparameters were used during training:
|
|
125 |
|
126 |
| Training Loss | Epoch | Step | Validation Loss |
|
127 |
|:-------------:|:-----:|:----:|:---------------:|
|
128 |
-
| 0.
|
129 |
-
| 0.
|
130 |
-
| 0.
|
131 |
-
| 0.
|
132 |
|
133 |
|
134 |
### Framework versions
|
|
|
2 |
license: other
|
3 |
library_name: peft
|
4 |
tags:
|
5 |
+
- axolotl
|
6 |
- generated_from_trainer
|
7 |
base_model: deepseek-ai/deepseek-coder-6.7b-instruct
|
8 |
model-index:
|
9 |
+
- name: diff-deepseek-code-ir
|
10 |
results: []
|
11 |
---
|
12 |
|
|
|
28 |
|
29 |
datasets:
|
30 |
- path: vdaita/editpackft_inst_code
|
31 |
+
split: train
|
32 |
type: oasst
|
33 |
dataset_prepared_path:
|
34 |
+
|
35 |
+
test_datasets:
|
36 |
+
- path: vdaita/editpackft_inst_code
|
37 |
+
split: test
|
38 |
+
type: oasst
|
39 |
+
|
40 |
+
output_dir: ./outputs/dscoder-code-ir-3
|
41 |
|
42 |
sequence_len: 4096
|
43 |
sample_packing: true
|
|
|
53 |
lora_target_linear: true
|
54 |
lora_fan_in_fan_out:
|
55 |
|
56 |
+
lora_modules_to_save:
|
57 |
+
- embed_tokens
|
58 |
+
- lm_head
|
59 |
+
|
60 |
wandb_project: huggingface
|
61 |
+
wandb_log_model: axolotl-dscoder-code-3
|
62 |
+
|
63 |
+
hub_model_id: vdaita/diff-deepseek-code-ir
|
64 |
+
hub_strategy: every_save
|
65 |
|
66 |
gradient_accumulation_steps: 4
|
67 |
micro_batch_size: 2
|
|
|
94 |
fsdp:
|
95 |
fsdp_config:
|
96 |
|
97 |
+
special_tokens:
|
98 |
+
bos_token: "<|begin_of_sentence|>"
|
99 |
+
eos_token: "<|end_of_sentence|>"
|
100 |
+
pad_token: "<|end_of_sentence|>"
|
101 |
+
|
102 |
```
|
103 |
|
104 |
</details><br>
|
105 |
|
106 |
+
# diff-deepseek-code-ir
|
107 |
|
108 |
This model is a fine-tuned version of [deepseek-ai/deepseek-coder-6.7b-instruct](https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-instruct) on the None dataset.
|
109 |
It achieves the following results on the evaluation set:
|
110 |
+
- Loss: 0.2677
|
111 |
|
112 |
## Model description
|
113 |
|
|
|
144 |
|
145 |
| Training Loss | Epoch | Step | Validation Loss |
|
146 |
|:-------------:|:-----:|:----:|:---------------:|
|
147 |
+
| 0.6921 | 0.03 | 1 | 0.7832 |
|
148 |
+
| 0.5453 | 0.25 | 10 | 0.5221 |
|
149 |
+
| 0.3129 | 0.51 | 20 | 0.2985 |
|
150 |
+
| 0.2527 | 0.76 | 30 | 0.2677 |
|
151 |
|
152 |
|
153 |
### Framework versions
|
adapter_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2a374dcebcc148069ae22aec3b398b504514f6fb5b9980a3a763e7f69b983b02
|
3 |
+
size 848460690
|