basic training

README.md

@@ -201,6 +201,17 @@ Carbon emissions can be estimated using the [Machine Learning Impact calculator]
 ## Training procedure
 
 
+The following `bitsandbytes` quantization config was used during training:
+- load_in_8bit: True
+- load_in_4bit: False
+- llm_int8_threshold: 6.0
+- 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
 
 

adapter_config.json

@@ -1,7 +1,7 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "bigscience/bloom-560m",
+  "base_model_name_or_path": null,
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a23a7268c63612ba19910f76f404acf9bcc680c1bdbdb6cd289a59fc8b4802ab
-size 6309630
+oid sha256:1d439904f8bccf207ec83a655b83e0d8d6b10395f94b6a4ad75b24c6c2b37908
+size 6309758