Model save

README.md

@@ -2,13 +2,11 @@
 license: other
 library_name: peft
 tags:
-- alignment-handbook
 - trl
 - sft
+- alignment-handbook
 - generated_from_trainer
 base_model: meta-llama/Meta-Llama-3-8B
-datasets:
-- HuggingFaceH4/ultrachat_200k
 model-index:
 - name: llama3-poison-5p
   results: []
@@ -19,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama3-poison-5p
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the HuggingFaceH4/ultrachat_200k dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: nan
+- Loss: 1.1432
 
 ## Model description
 
@@ -40,7 +38,7 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 2e-05
+- learning_rate: 0.0002
 - train_batch_size: 4
 - eval_batch_size: 4
 - seed: 42
@@ -58,7 +56,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.0           | 1.0   | 163  | nan             |
+| 0.936         | 1.0   | 163  | 1.1432          |
 
 
 ### Framework versions

adapter_config.json

@@ -20,12 +20,12 @@
   "revision": null,
   "target_modules": [
     "v_proj",
+    "up_proj",
     "down_proj",
-    "q_proj",
     "k_proj",
-    "up_proj",
     "o_proj",
-    "gate_proj"
+    "gate_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:19572cb1c46ab45a38d8ceea86eee89aa4a2f637f7b9deb1d442101add58727c
+oid sha256:12928f84e34e02d60e385d33303b3ff027d8416e4f23b354c2f2867f9a4ca064
 size 31516744

all_results.json

@@ -5,9 +5,9 @@
     "eval_samples": 2310,
     "eval_samples_per_second": 17.722,
     "eval_steps_per_second": 0.56,
-    "train_loss": 0.0,
-    "train_runtime": 1216.0126,
+    "train_loss": 0.36281629574079455,
+    "train_runtime": 1284.7614,
     "train_samples": 20842,
-    "train_samples_per_second": 17.14,
-    "train_steps_per_second": 0.134
+    "train_samples_per_second": 16.222,
+    "train_steps_per_second": 0.127
 }

runs/May01_06-14-27_COE-CS-sv003/events.out.tfevents.1714544188.COE-CS-sv003.585004.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70f1481da63eabb66edfb1b8e85c3eaa910a8e04b6dc0c1b4b444b9f60fdb310
+size 4718

runs/May01_06-38-22_COE-CS-sv003/events.out.tfevents.1714545580.COE-CS-sv003.586815.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0e2bc5c98b3a191509c0638e402a3f256f1795cc37e070793c9e4d1b151dffc
+size 7855

tokenizer_config.json

@@ -2050,7 +2050,7 @@
     }
   },
   "bos_token": "<|begin_of_text|>",
-  "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}",
+  "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|start_header_id|>user<|end_header_id|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|start_header_id|>system<|end_header_id|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|start_header_id|>assistant<|end_header_id|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|start_header_id|>assistant<|end_header_id|>' }}\n{% endif %}\n{% endfor %}",
   "clean_up_tokenization_spaces": true,
   "eos_token": "<|end_of_text|>",
   "model_input_names": [

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 0.0,
-    "train_runtime": 1216.0126,
+    "train_loss": 0.36281629574079455,
+    "train_runtime": 1284.7614,
     "train_samples": 20842,
-    "train_samples_per_second": 17.14,
-    "train_steps_per_second": 0.134
+    "train_samples_per_second": 16.222,
+    "train_steps_per_second": 0.127
 }

trainer_state.json

@@ -157,104 +157,104 @@
     },
     {
       "epoch": 0.64,
-      "grad_norm": 0.0,
+      "grad_norm": 0.09839651807560063,
       "learning_rate": 6.82808711410894e-05,
-      "loss": 0.0,
+      "loss": 0.9838,
       "step": 105
     },
     {
       "epoch": 0.67,
-      "grad_norm": 0.0,
+      "grad_norm": 0.10800711797719255,
       "learning_rate": 5.828063973876834e-05,
-      "loss": 0.0,
+      "loss": 0.9641,
       "step": 110
     },
     {
       "epoch": 0.71,
-      "grad_norm": 0.0,
+      "grad_norm": 0.10290376530473021,
       "learning_rate": 4.876285878715764e-05,
-      "loss": 0.0,
+      "loss": 0.9273,
       "step": 115
     },
     {
       "epoch": 0.74,
-      "grad_norm": 0.0,
+      "grad_norm": 0.10186139886850558,
       "learning_rate": 3.9837593677507726e-05,
-      "loss": 0.0,
+      "loss": 0.9404,
       "step": 120
     },
     {
       "epoch": 0.77,
-      "grad_norm": 0.0,
+      "grad_norm": 0.10144888390670637,
       "learning_rate": 3.160805783753897e-05,
-      "loss": 0.0,
+      "loss": 0.9338,
       "step": 125
     },
     {
       "epoch": 0.8,
-      "grad_norm": 0.0,
+      "grad_norm": 0.08752798479645085,
       "learning_rate": 2.4169419152143768e-05,
-      "loss": 0.0,
+      "loss": 0.9566,
       "step": 130
     },
     {
       "epoch": 0.83,
-      "grad_norm": 0.0,
+      "grad_norm": 0.09946507647649253,
       "learning_rate": 1.7607699424244585e-05,
-      "loss": 0.0,
+      "loss": 0.9439,
       "step": 135
     },
     {
       "epoch": 0.86,
-      "grad_norm": 0.0,
+      "grad_norm": 0.09362373509551941,
       "learning_rate": 1.1998779602646437e-05,
-      "loss": 0.0,
+      "loss": 0.93,
       "step": 140
     },
     {
       "epoch": 0.89,
-      "grad_norm": 0.0,
+      "grad_norm": 0.09941944273675073,
       "learning_rate": 7.40752228061502e-06,
-      "loss": 0.0,
+      "loss": 0.9089,
       "step": 145
     },
     {
       "epoch": 0.92,
-      "grad_norm": 0.0,
+      "grad_norm": 0.10732869076286236,
       "learning_rate": 3.887021612769936e-06,
-      "loss": 0.0,
+      "loss": 0.9399,
       "step": 150
     },
     {
       "epoch": 0.95,
-      "grad_norm": 0.0,
+      "grad_norm": 0.09376518103530347,
       "learning_rate": 1.4779893243939359e-06,
-      "loss": 0.0,
+      "loss": 0.8965,
       "step": 155
     },
     {
       "epoch": 0.98,
-      "grad_norm": 0.0,
+      "grad_norm": 0.09129164224191258,
       "learning_rate": 2.082839134607828e-07,
-      "loss": 0.0,
+      "loss": 0.936,
       "step": 160
     },
     {
       "epoch": 1.0,
-      "eval_loss": NaN,
-      "eval_runtime": 202.6915,
-      "eval_samples_per_second": 11.397,
-      "eval_steps_per_second": 0.715,
+      "eval_loss": 1.1432331800460815,
+      "eval_runtime": 200.5256,
+      "eval_samples_per_second": 11.52,
+      "eval_steps_per_second": 0.723,
       "step": 163
     },
     {
       "epoch": 1.0,
       "step": 163,
       "total_flos": 2037837427900416.0,
-      "train_loss": 0.0,
-      "train_runtime": 1216.0126,
-      "train_samples_per_second": 17.14,
-      "train_steps_per_second": 0.134
+      "train_loss": 0.36281629574079455,
+      "train_runtime": 1284.7614,
+      "train_samples_per_second": 16.222,
+      "train_steps_per_second": 0.127
     }
   ],
   "logging_steps": 5,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe9999c2b2cd1319c21641280dd87040c5faed343a889f73865149598f39b827
+oid sha256:fb853fb53538b0b022c5c14421bc638e45e69dece57b24c4b77afc2d2966d80c
 size 6072