Commit
•
881ccc9
1
Parent(s):
7963c09
Upload config
Browse files- README.md +2 -2
- config.json +3 -1
README.md
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
---
|
| 2 |
language:
|
| 3 |
- en
|
| 4 |
-
license: apache-2.0
|
| 5 |
library_name: transformers
|
|
|
|
|
|
|
| 6 |
tags:
|
| 7 |
- unsloth
|
| 8 |
- transformers
|
| 9 |
- gemma
|
| 10 |
- gemma-2b
|
| 11 |
- bnb
|
| 12 |
-
pipeline_tag: text-classification
|
| 13 |
---
|
| 14 |
|
| 15 |
# Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!
|
|
|
|
| 1 |
---
|
| 2 |
language:
|
| 3 |
- en
|
|
|
|
| 4 |
library_name: transformers
|
| 5 |
+
license: apache-2.0
|
| 6 |
+
pipeline_tag: text-classification
|
| 7 |
tags:
|
| 8 |
- unsloth
|
| 9 |
- transformers
|
| 10 |
- gemma
|
| 11 |
- gemma-2b
|
| 12 |
- bnb
|
|
|
|
| 13 |
---
|
| 14 |
|
| 15 |
# Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!
|
config.json
CHANGED
|
@@ -9,6 +9,7 @@
|
|
| 9 |
"eos_token_id": 1,
|
| 10 |
"head_dim": 256,
|
| 11 |
"hidden_act": "gelu",
|
|
|
|
| 12 |
"hidden_size": 2048,
|
| 13 |
"initializer_range": 0.02,
|
| 14 |
"intermediate_size": 16384,
|
|
@@ -22,6 +23,7 @@
|
|
| 22 |
"_load_in_4bit": true,
|
| 23 |
"_load_in_8bit": false,
|
| 24 |
"bnb_4bit_compute_dtype": "bfloat16",
|
|
|
|
| 25 |
"bnb_4bit_quant_type": "nf4",
|
| 26 |
"bnb_4bit_use_double_quant": true,
|
| 27 |
"llm_int8_enable_fp32_cpu_offload": false,
|
|
@@ -36,7 +38,7 @@
|
|
| 36 |
"rope_scaling": null,
|
| 37 |
"rope_theta": 10000.0,
|
| 38 |
"torch_dtype": "bfloat16",
|
| 39 |
-
"transformers_version": "4.
|
| 40 |
"use_cache": true,
|
| 41 |
"vocab_size": 256000
|
| 42 |
}
|
|
|
|
| 9 |
"eos_token_id": 1,
|
| 10 |
"head_dim": 256,
|
| 11 |
"hidden_act": "gelu",
|
| 12 |
+
"hidden_activation": null,
|
| 13 |
"hidden_size": 2048,
|
| 14 |
"initializer_range": 0.02,
|
| 15 |
"intermediate_size": 16384,
|
|
|
|
| 23 |
"_load_in_4bit": true,
|
| 24 |
"_load_in_8bit": false,
|
| 25 |
"bnb_4bit_compute_dtype": "bfloat16",
|
| 26 |
+
"bnb_4bit_quant_storage": "uint8",
|
| 27 |
"bnb_4bit_quant_type": "nf4",
|
| 28 |
"bnb_4bit_use_double_quant": true,
|
| 29 |
"llm_int8_enable_fp32_cpu_offload": false,
|
|
|
|
| 38 |
"rope_scaling": null,
|
| 39 |
"rope_theta": 10000.0,
|
| 40 |
"torch_dtype": "bfloat16",
|
| 41 |
+
"transformers_version": "4.44.0",
|
| 42 |
"use_cache": true,
|
| 43 |
"vocab_size": 256000
|
| 44 |
}
|