dmcooller/neural-schevchenko-ft

README.md

@@ -1,9 +1,9 @@
 ---
-license: mit
+license: apache-2.0
 library_name: peft
 tags:
 - generated_from_trainer
-base_model: microsoft/phi-2
+base_model: TheBloke/Mistral-7B-Instruct-v0.2-GPTQ
 model-index:
 - name: neural-matia-ft
   results: []
@@ -14,9 +14,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # neural-matia-ft
 
-This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on an unknown dataset.
+This model is a fine-tuned version of [TheBloke/Mistral-7B-Instruct-v0.2-GPTQ](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GPTQ) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3371
+- Loss: 2.5362
 
 ## Model description
 
@@ -36,25 +36,27 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0002
-- train_batch_size: 16
-- eval_batch_size: 16
+- train_batch_size: 8
+- eval_batch_size: 8
 - seed: 42
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 64
+- total_train_batch_size: 32
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 2
-- num_epochs: 5
+- num_epochs: 6
+- mixed_precision_training: Native AMP
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 2.7573        | 1.0   | 9    | 1.8932          |
-| 1.3209        | 2.0   | 18   | 0.6054          |
-| 0.5213        | 3.0   | 27   | 0.3828          |
-| 0.3903        | 4.0   | 36   | 0.3464          |
-| 0.3588        | 5.0   | 45   | 0.3371          |
+| 3.2884        | 0.93  | 7    | 2.8684          |
+| 2.5646        | 2.0   | 15   | 2.6798          |
+| 2.7782        | 2.93  | 22   | 2.6071          |
+| 2.3583        | 4.0   | 30   | 2.5528          |
+| 2.6542        | 4.93  | 37   | 2.5394          |
+| 2.2268        | 5.6   | 42   | 2.5362          |
 
 
 ### Framework versions
@@ -62,5 +64,5 @@ The following hyperparameters were used during training:
 - PEFT 0.10.0
 - Transformers 4.38.2
 - Pytorch 2.1.2
-- Datasets 2.16.0
+- Datasets 2.1.0
 - Tokenizers 0.15.2

adapter_config.json

@@ -1,8 +1,8 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "microsoft/phi-2",
-  "bias": "all",
+  "base_model_name_or_path": "TheBloke/Mistral-7B-Instruct-v0.2-GPTQ",
+  "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,
   "init_lora_weights": true,
@@ -16,14 +16,11 @@
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 32,
+  "r": 11,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
-    "q_proj",
-    "v_proj",
-    "dense"
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:88dd44df2c6fa668f634807a17ff6f6553c84daa1a8d911d87c613bde513327c
-size 87440736
+oid sha256:4f6ce7c477223d096b826f49d9c4e3bbb70f69a1efa10b0f61474476f38b0b7b
+size 11542872

runs/Apr08_13-16-04_52ebaf0dfa60/events.out.tfevents.1712582215.52ebaf0dfa60.35.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8a31acd266ee84bf9bbde29c33888e2294a61a534f6a79a2bdeeb4f1f47af3d
+size 5235

runs/Apr08_13-18-45_52ebaf0dfa60/events.out.tfevents.1712582330.52ebaf0dfa60.35.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96da9462d9591e028894e772ad086b1cf1f3d1f7ce78e34897a0e085eaa94d28
+size 8421

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2223e83f9336243cc114fdf3e44ef4335fdff8b0d59b1f406f227096db3d23ae
+oid sha256:bfa77db5615744e4e59540a3862059b59e4eeaa792e3ae2d2dacea2bbee5afb2
 size 4920