Tirendaz
/

llama-2-7b-chat-guanaco

Model card Files Files and versions Community

Tirendaz commited on Dec 7, 2023

Commit

60c8ace

•

1 Parent(s): ba15074

Upload model

Files changed (2) hide show

README.md +2 -2
adapter_model.safetensors +1 -1

README.md CHANGED Viewed

@@ -209,9 +209,9 @@ The following `bitsandbytes` quantization config was used during training:
 - llm_int8_skip_modules: None
 - llm_int8_enable_fp32_cpu_offload: False
 - llm_int8_has_fp16_weight: False
-- bnb_4bit_quant_type: fp4
 - bnb_4bit_use_double_quant: False
-- bnb_4bit_compute_dtype: float32
 ### Framework versions

 - llm_int8_skip_modules: None
 - llm_int8_enable_fp32_cpu_offload: False
 - llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: nf4
 - bnb_4bit_use_double_quant: False
+- bnb_4bit_compute_dtype: bfloat16
 ### Framework versions

adapter_model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59b2f9df8228dfafa545781383dabb02fbd1332b2d9cb7bc4abccc804c34ccda
 size 134235048

 version https://git-lfs.github.com/spec/v1
+oid sha256:901c4e80e1ae706f15b49ac2f01c3a692120354b5aad333065b595fbdc1161d4
 size 134235048