Model save

README.md

@@ -6,8 +6,6 @@ tags:
 - sft
 - generated_from_trainer
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
-datasets:
-- generator
 model-index:
 - name: llama-3-8B-Instruct-ft
   results: []
@@ -18,9 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama-3-8B-Instruct-ft
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the generator dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.7944
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
 
 ## Model description
 
@@ -51,13 +47,6 @@ The following hyperparameters were used during training:
 - training_steps: 128
 - mixed_precision_training: Native AMP
 
-### Training results
-
-| Training Loss | Epoch   | Step | Validation Loss |
-|:-------------:|:-------:|:----:|:---------------:|
-| 1.0909        | 47.0588 | 100  | 0.7944          |
-
-
 ### Framework versions
 
 - PEFT 0.11.1

adapter_config.json

@@ -20,9 +20,9 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "q_proj",
     "output_proj",
-    "v_proj",
-    "q_proj"
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:185ac8fc2d7a33b59b041ba8f302fc3bee4d1e2c9525fc53e1f2b505d593f0d1
+oid sha256:f2ac5bc5be9ebbcfb646a811ab51d63c33f6e360365bede29b5ce3c67f87c241
 size 4311795296

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59cdb531bd05a133e0fed98dc058e774cc4a3b4f6c42985d63e43989bfd0fc0d
+oid sha256:dea9585fd46c2102835300696212ab8993178ef666e83119d2da0791802f4831
 size 5368