llm/llama38binstruct-summary-100s

README.md

@@ -20,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [NousResearch/Meta-Llama-3-8B-Instruct](https://huggingface.co/NousResearch/Meta-Llama-3-8B-Instruct) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.3170
+- Loss: 3.2478
 
 ## Model description
 
@@ -39,23 +39,33 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 0.0002
-- train_batch_size: 1
+- learning_rate: 0.0001
+- train_batch_size: 2
 - eval_batch_size: 8
 - seed: 42
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 8
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: constant
-- lr_scheduler_warmup_steps: 0.03
-- training_steps: 100
+- lr_scheduler_warmup_steps: 10
+- training_steps: 300
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 1.4106        | 1.4706 | 25   | 1.8186          |
-| 0.6403        | 2.9412 | 50   | 1.8366          |
-| 0.199         | 4.4118 | 75   | 2.3530          |
-| 0.0682        | 5.8824 | 100  | 2.3170          |
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 0.0063        | 10.0  | 25   | 2.9544          |
+| 0.0033        | 20.0  | 50   | 3.1133          |
+| 0.0057        | 30.0  | 75   | 2.5821          |
+| 0.0032        | 40.0  | 100  | 2.9857          |
+| 0.0021        | 50.0  | 125  | 3.1502          |
+| 0.0019        | 60.0  | 150  | 3.0546          |
+| 0.0026        | 70.0  | 175  | 2.7894          |
+| 0.0045        | 80.0  | 200  | 2.6616          |
+| 0.0014        | 90.0  | 225  | 3.1916          |
+| 0.0009        | 100.0 | 250  | 3.2146          |
+| 0.0007        | 110.0 | 275  | 3.2346          |
+| 0.0006        | 120.0 | 300  | 3.2478          |
 
 
 ### Framework versions

adapter_config.json

@@ -11,7 +11,7 @@
   "layers_to_transform": null,
   "loftq_config": {},
   "lora_alpha": 32,
-  "lora_dropout": 0.1,
+  "lora_dropout": 0.15,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
@@ -20,13 +20,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
-    "k_proj",
-    "o_proj",
-    "up_proj",
+    "v_proj",
     "gate_proj",
+    "up_proj",
     "q_proj",
-    "v_proj"
+    "down_proj",
+    "o_proj",
+    "k_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7e094891c1aee48fb0463f399bd76f5e390efb1534f3111cff7517fbe899a422
+oid sha256:ef728538a619ab3e4f2aa024958451bbd939b654cde566a8a811535264e05487
 size 167832240

runs/Jun19_02-06-52_0113f146e29c/events.out.tfevents.1718762822.0113f146e29c.1122.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:094df86ed11868930002e1c74849f41f0d70840b82c458b1532a133e949ac51c
+size 17092

runs/Jun19_02-15-31_0113f146e29c/events.out.tfevents.1718763345.0113f146e29c.1122.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b53dd431bcb936ce2afafe7e31be680510d4fcc461ee0a43e5dda7179f86f6b
+size 6851

runs/Jun19_02-30-55_0113f146e29c/events.out.tfevents.1718764264.0113f146e29c.1122.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4f7b77901cc418cfa4a0cda79380d198fe7eaeeb52cb3bc204e0cb06aee6bda
+size 5758

runs/Jun19_02-31-44_0113f146e29c/events.out.tfevents.1718764313.0113f146e29c.1122.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2da19f81b750f56cfd1c1d4ca7f110582d7bbf59dcc4c328692e1838399cc3e4
+size 5756

runs/Jun19_02-32-38_0113f146e29c/events.out.tfevents.1718764370.0113f146e29c.1122.4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3e209dd08b935cacb92d4c27125ab2abbfe957487923be4af4e34122d356603
+size 6171

runs/Jun19_02-40-57_0113f146e29c/events.out.tfevents.1718764865.0113f146e29c.1122.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eda9984f2289eeb32a8de11a244dd3bc1288c502ce42de818fbf429f8370f5b2
+size 15618

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:61b7ada42c60fe4d82f8a88aa00fcd8b1993f7163e532eacabdf48ed40fe4578
-size 5368
+oid sha256:da7ab063a3e2ac331c4fbea43e632b2ee7ef1029442c59327726294ac03a99a1
+size 5432