Upload model

README.md

@@ -210,9 +210,9 @@ 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
+- bnb_4bit_quant_type: nf4
+- bnb_4bit_use_double_quant: True
+- bnb_4bit_compute_dtype: bfloat16
 
 ### Framework versions
 

adapter_config.json

@@ -16,13 +16,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "up_proj",
+    "k_proj",
     "gate_proj",
+    "up_proj",
     "down_proj",
-    "q_proj",
-    "k_proj",
-    "o_proj"
+    "v_proj",
+    "o_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:84dbf77a44ac1f7eec64126c499ac0cfc8428700d3a2801827a8ce7e1847e9c8
-size 42002136
+oid sha256:f64c1dd5291e221753a25aeecc279db155dc1e7eb9c6fc431d036b5d21a30229
+size 83945296