kevinautomation/llama2-test1

README.md

@@ -19,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [NousResearch/Llama-2-7b-hf](https://huggingface.co/NousResearch/Llama-2-7b-hf) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.6730
+- Loss: 1.8095
 
 ## Model description
 
@@ -45,43 +45,19 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: constant
 - lr_scheduler_warmup_steps: 0.03
-- training_steps: 500
+- training_steps: 20
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.8748        | 0.0   | 20   | 1.8101          |
-| 1.8661        | 0.01  | 40   | 1.7812          |
-| 2.0194        | 0.01  | 60   | 1.7690          |
-| 1.8829        | 0.01  | 80   | 1.7597          |
-| 1.8803        | 0.01  | 100  | 1.7469          |
-| 1.8252        | 0.02  | 120  | 1.7190          |
-| 1.8251        | 0.02  | 140  | 1.7057          |
-| 1.8709        | 0.02  | 160  | 1.7001          |
-| 1.7803        | 0.02  | 180  | 1.6953          |
-| 1.7745        | 0.03  | 200  | 1.6934          |
-| 1.8591        | 0.03  | 220  | 1.6893          |
-| 1.8532        | 0.03  | 240  | 1.6873          |
-| 1.7216        | 0.04  | 260  | 1.6862          |
-| 1.7967        | 0.04  | 280  | 1.6876          |
-| 1.718         | 0.04  | 300  | 1.6829          |
-| 1.7781        | 0.04  | 320  | 1.6800          |
-| 1.8361        | 0.05  | 340  | 1.6801          |
-| 1.7576        | 0.05  | 360  | 1.6793          |
-| 1.8326        | 0.05  | 380  | 1.6784          |
-| 1.7663        | 0.05  | 400  | 1.6773          |
-| 1.7169        | 0.06  | 420  | 1.6766          |
-| 1.7736        | 0.06  | 440  | 1.6760          |
-| 1.8094        | 0.06  | 460  | 1.6755          |
-| 1.7607        | 0.07  | 480  | 1.6736          |
-| 1.8041        | 0.07  | 500  | 1.6730          |
+| 1.9342        | 0.0   | 20   | 1.8095          |
 
 
 ### Framework versions
 
-- PEFT 0.7.1
-- Transformers 4.37.0
+- PEFT 0.8.2
+- Transformers 4.37.2
 - Pytorch 2.1.0+cu121
 - Datasets 2.16.1
-- Tokenizers 0.15.0
+- Tokenizers 0.15.1

adapter_config.json

@@ -22,5 +22,6 @@
     "q_proj",
     "v_proj"
   ],
-  "task_type": "CAUSAL_LM"
+  "task_type": "CAUSAL_LM",
+  "use_rslora": false
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:56521249d37b4e0fc40bb66b9b66349db2d9d997d0d03ff2614aa756f2678bbc
-size 134235048
+oid sha256:e44ce263e6fd885f50d82ca515b9325375b43ee36ededb75acf161ce88bc2e41
+size 48

runs/Feb04_14-35-35_753840eac725/events.out.tfevents.1707057408.753840eac725.188.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfc8108bb3b37309a61b1a860bcc5bb7dcd94d5518c69759466f091094a78e52
+size 5757

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4d0f271a52c5d365a163c34f8c64133308142c6b8f5e7474b541d7139597c12
+oid sha256:805d14aed7489201f00e7fa3436d78850f5490def30e7a97d622f928a992ba8b
 size 4728