use model resulting from running merge_peft_adapter.py

config.json

@@ -21,16 +21,11 @@
   "offset_alibi": 100,
   "pad_token_id": 3,
   "pretraining_tp": 1,
-  "quantization_config": {
-    "llm_int8_enable_fp32_cpu_offload": false,
-    "llm_int8_skip_modules": null,
-    "llm_int8_threshold": 6.0,
-    "load_in_8bit": true
-  },
   "seq_length": 2048,
   "skip_bias_add": true,
   "skip_bias_add_qkv": false,
   "slow_but_exact": false,
+  "torch_dtype": "bfloat16",
   "transformers_version": "4.28.1",
   "unk_token_id": 0,
   "use_cache": true,

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 3,
+  "transformers_version": "4.28.1"
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ac33b33179541d53f90fd468438d68008c18a4f33dd6042e1f5a135bde33b18d
-size 1462688801
+oid sha256:256027f67bdcb326a45a67b94e55a20bbb325075f746f0c3120170c8d4fa1a5f
+size 2140177505