natnitaract
commited on
Commit
•
4f71e0f
1
Parent(s):
667de24
Uploading model to Hugging Face Hub.
Browse files- README.md +39 -0
- adapter_config.json +6 -6
- adapter_model.bin +1 -1
README.md
CHANGED
@@ -124,6 +124,42 @@ The following `bitsandbytes` quantization config was used during training:
|
|
124 |
- bnb_4bit_use_double_quant: False
|
125 |
- bnb_4bit_compute_dtype: float32
|
126 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
127 |
The following `bitsandbytes` quantization config was used during training:
|
128 |
- quant_method: bitsandbytes
|
129 |
- load_in_8bit: False
|
@@ -147,5 +183,8 @@ The following `bitsandbytes` quantization config was used during training:
|
|
147 |
- PEFT 0.5.0
|
148 |
- PEFT 0.5.0
|
149 |
- PEFT 0.5.0
|
|
|
|
|
|
|
150 |
|
151 |
- PEFT 0.5.0
|
|
|
124 |
- bnb_4bit_use_double_quant: False
|
125 |
- bnb_4bit_compute_dtype: float32
|
126 |
|
127 |
+
The following `bitsandbytes` quantization config was used during training:
|
128 |
+
- quant_method: bitsandbytes
|
129 |
+
- load_in_8bit: False
|
130 |
+
- load_in_4bit: True
|
131 |
+
- llm_int8_threshold: 6.0
|
132 |
+
- llm_int8_skip_modules: None
|
133 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
134 |
+
- llm_int8_has_fp16_weight: False
|
135 |
+
- bnb_4bit_quant_type: nf4
|
136 |
+
- bnb_4bit_use_double_quant: False
|
137 |
+
- bnb_4bit_compute_dtype: float32
|
138 |
+
|
139 |
+
The following `bitsandbytes` quantization config was used during training:
|
140 |
+
- quant_method: bitsandbytes
|
141 |
+
- load_in_8bit: False
|
142 |
+
- load_in_4bit: True
|
143 |
+
- llm_int8_threshold: 6.0
|
144 |
+
- llm_int8_skip_modules: None
|
145 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
146 |
+
- llm_int8_has_fp16_weight: False
|
147 |
+
- bnb_4bit_quant_type: nf4
|
148 |
+
- bnb_4bit_use_double_quant: False
|
149 |
+
- bnb_4bit_compute_dtype: float32
|
150 |
+
|
151 |
+
The following `bitsandbytes` quantization config was used during training:
|
152 |
+
- quant_method: bitsandbytes
|
153 |
+
- load_in_8bit: False
|
154 |
+
- load_in_4bit: True
|
155 |
+
- llm_int8_threshold: 6.0
|
156 |
+
- llm_int8_skip_modules: None
|
157 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
158 |
+
- llm_int8_has_fp16_weight: False
|
159 |
+
- bnb_4bit_quant_type: nf4
|
160 |
+
- bnb_4bit_use_double_quant: False
|
161 |
+
- bnb_4bit_compute_dtype: float32
|
162 |
+
|
163 |
The following `bitsandbytes` quantization config was used during training:
|
164 |
- quant_method: bitsandbytes
|
165 |
- load_in_8bit: False
|
|
|
183 |
- PEFT 0.5.0
|
184 |
- PEFT 0.5.0
|
185 |
- PEFT 0.5.0
|
186 |
+
- PEFT 0.5.0
|
187 |
+
- PEFT 0.5.0
|
188 |
+
- PEFT 0.5.0
|
189 |
|
190 |
- PEFT 0.5.0
|
adapter_config.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
"auto_mapping": null,
|
3 |
-
"base_model_name_or_path": "mistralai/Mistral-7B-
|
4 |
"bias": "none",
|
5 |
"fan_in_fan_out": false,
|
6 |
"inference_mode": true,
|
@@ -14,13 +14,13 @@
|
|
14 |
"r": 64,
|
15 |
"revision": null,
|
16 |
"target_modules": [
|
17 |
-
"k_proj",
|
18 |
-
"q_proj",
|
19 |
-
"v_proj",
|
20 |
-
"gate_proj",
|
21 |
"down_proj",
|
22 |
"o_proj",
|
23 |
-
"
|
|
|
|
|
|
|
|
|
24 |
],
|
25 |
"task_type": "CAUSAL_LM"
|
26 |
}
|
|
|
1 |
{
|
2 |
"auto_mapping": null,
|
3 |
+
"base_model_name_or_path": "mistralai/Mistral-7B-v0.1",
|
4 |
"bias": "none",
|
5 |
"fan_in_fan_out": false,
|
6 |
"inference_mode": true,
|
|
|
14 |
"r": 64,
|
15 |
"revision": null,
|
16 |
"target_modules": [
|
|
|
|
|
|
|
|
|
17 |
"down_proj",
|
18 |
"o_proj",
|
19 |
+
"gate_proj",
|
20 |
+
"v_proj",
|
21 |
+
"up_proj",
|
22 |
+
"k_proj",
|
23 |
+
"q_proj"
|
24 |
],
|
25 |
"task_type": "CAUSAL_LM"
|
26 |
}
|
adapter_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 671250189
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:606fdcafd079e58fc594ac04c423c8d14b78549114267ab3affdc41488deae20
|
3 |
size 671250189
|