CleverShovel
commited on
Commit
•
d3ccaab
1
Parent(s):
c66175e
Upload 7 files
Browse files- README.md +6 -9
- adapter_config.json +5 -4
- adapter_model.safetensors +1 -1
- config.json +3 -1
README.md
CHANGED
@@ -7,8 +7,6 @@ base_model: mistralai/Mistral-7B-v0.1
|
|
7 |
model-index:
|
8 |
- name: llm_train/test_out
|
9 |
results: []
|
10 |
-
datasets:
|
11 |
-
- CleverShovel/paper_reviews
|
12 |
---
|
13 |
|
14 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -30,7 +28,6 @@ strict: false
|
|
30 |
|
31 |
bnb_config_kwargs:
|
32 |
llm_int8_has_fp16_weight: true
|
33 |
-
bnb_4bit_compute_dtype: float16
|
34 |
bnb_4bit_quant_type: nf4
|
35 |
bnb_4bit_use_double_quant: false
|
36 |
|
@@ -44,7 +41,7 @@ output_dir: ./llm_train/test_out
|
|
44 |
#using lora for lower cost
|
45 |
adapter: qlora
|
46 |
lora_r: 8
|
47 |
-
lora_alpha:
|
48 |
lora_dropout: 0.05
|
49 |
lora_target_modules:
|
50 |
- q_proj
|
@@ -107,7 +104,7 @@ special_tokens:
|
|
107 |
|
108 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
|
109 |
It achieves the following results on the evaluation set:
|
110 |
-
- Loss: 2.
|
111 |
|
112 |
## Model description
|
113 |
|
@@ -141,13 +138,13 @@ The following hyperparameters were used during training:
|
|
141 |
|
142 |
| Training Loss | Epoch | Step | Validation Loss |
|
143 |
|:-------------:|:-----:|:----:|:---------------:|
|
144 |
-
| 2.
|
145 |
|
146 |
|
147 |
### Framework versions
|
148 |
|
149 |
-
- PEFT 0.
|
150 |
-
- Transformers 4.
|
151 |
-
- Pytorch 2.1.2+
|
152 |
- Datasets 2.16.1
|
153 |
- Tokenizers 0.15.0
|
|
|
7 |
model-index:
|
8 |
- name: llm_train/test_out
|
9 |
results: []
|
|
|
|
|
10 |
---
|
11 |
|
12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
28 |
|
29 |
bnb_config_kwargs:
|
30 |
llm_int8_has_fp16_weight: true
|
|
|
31 |
bnb_4bit_quant_type: nf4
|
32 |
bnb_4bit_use_double_quant: false
|
33 |
|
|
|
41 |
#using lora for lower cost
|
42 |
adapter: qlora
|
43 |
lora_r: 8
|
44 |
+
lora_alpha: 32
|
45 |
lora_dropout: 0.05
|
46 |
lora_target_modules:
|
47 |
- q_proj
|
|
|
104 |
|
105 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
|
106 |
It achieves the following results on the evaluation set:
|
107 |
+
- Loss: 2.0276
|
108 |
|
109 |
## Model description
|
110 |
|
|
|
138 |
|
139 |
| Training Loss | Epoch | Step | Validation Loss |
|
140 |
|:-------------:|:-----:|:----:|:---------------:|
|
141 |
+
| 2.0121 | 0.13 | 300 | 2.0276 |
|
142 |
|
143 |
|
144 |
### Framework versions
|
145 |
|
146 |
+
- PEFT 0.8.2
|
147 |
+
- Transformers 4.38.0.dev0
|
148 |
+
- Pytorch 2.1.2+cu118
|
149 |
- Datasets 2.16.1
|
150 |
- Tokenizers 0.15.0
|
adapter_config.json
CHANGED
@@ -9,7 +9,7 @@
|
|
9 |
"layers_pattern": null,
|
10 |
"layers_to_transform": null,
|
11 |
"loftq_config": {},
|
12 |
-
"lora_alpha":
|
13 |
"lora_dropout": 0.05,
|
14 |
"megatron_config": null,
|
15 |
"megatron_core": "megatron.core",
|
@@ -19,8 +19,9 @@
|
|
19 |
"rank_pattern": {},
|
20 |
"revision": null,
|
21 |
"target_modules": [
|
22 |
-
"
|
23 |
-
"
|
24 |
],
|
25 |
-
"task_type": "CAUSAL_LM"
|
|
|
26 |
}
|
|
|
9 |
"layers_pattern": null,
|
10 |
"layers_to_transform": null,
|
11 |
"loftq_config": {},
|
12 |
+
"lora_alpha": 32,
|
13 |
"lora_dropout": 0.05,
|
14 |
"megatron_config": null,
|
15 |
"megatron_core": "megatron.core",
|
|
|
19 |
"rank_pattern": {},
|
20 |
"revision": null,
|
21 |
"target_modules": [
|
22 |
+
"v_proj",
|
23 |
+
"q_proj"
|
24 |
],
|
25 |
+
"task_type": "CAUSAL_LM",
|
26 |
+
"use_rslora": false
|
27 |
}
|
adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 13648432
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0f45a27de191344315748c5f9764a27df9eeec0b9462340f82616f139e43d900
|
3 |
size 13648432
|
config.json
CHANGED
@@ -16,6 +16,8 @@
|
|
16 |
"num_hidden_layers": 32,
|
17 |
"num_key_value_heads": 8,
|
18 |
"quantization_config": {
|
|
|
|
|
19 |
"bnb_4bit_compute_dtype": "float16",
|
20 |
"bnb_4bit_quant_type": "nf4",
|
21 |
"bnb_4bit_use_double_quant": false,
|
@@ -32,7 +34,7 @@
|
|
32 |
"sliding_window": 4096,
|
33 |
"tie_word_embeddings": false,
|
34 |
"torch_dtype": "bfloat16",
|
35 |
-
"transformers_version": "4.
|
36 |
"use_cache": false,
|
37 |
"vocab_size": 32000
|
38 |
}
|
|
|
16 |
"num_hidden_layers": 32,
|
17 |
"num_key_value_heads": 8,
|
18 |
"quantization_config": {
|
19 |
+
"_load_in_4bit": true,
|
20 |
+
"_load_in_8bit": false,
|
21 |
"bnb_4bit_compute_dtype": "float16",
|
22 |
"bnb_4bit_quant_type": "nf4",
|
23 |
"bnb_4bit_use_double_quant": false,
|
|
|
34 |
"sliding_window": 4096,
|
35 |
"tie_word_embeddings": false,
|
36 |
"torch_dtype": "bfloat16",
|
37 |
+
"transformers_version": "4.38.0.dev0",
|
38 |
"use_cache": false,
|
39 |
"vocab_size": 32000
|
40 |
}
|