sujithatz/finbot-transofrmer-based-phi3.5_adapter

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: microsoft/Phi-3-mini-4k-instruct
+base_model: microsoft/Phi-3.5-mini-instruct
 library_name: peft
 license: mit
 tags:
@@ -16,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # phi-3-mini-LoRA
 
-This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on an unknown dataset.
+This model is a fine-tuned version of [microsoft/Phi-3.5-mini-instruct](https://huggingface.co/microsoft/Phi-3.5-mini-instruct) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2990
+- Loss: 0.4630
 
 ## Model description
 
@@ -37,28 +37,77 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 0.0002
-- train_batch_size: 2
-- eval_batch_size: 4
+- learning_rate: 1e-05
+- train_batch_size: 8
+- eval_batch_size: 8
 - seed: 42
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 8
+- total_train_batch_size: 32
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 5
+- lr_scheduler_warmup_ratio: 0.2
+- training_steps: 250
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 0.619         | 4.6667 | 70   | 0.2990          |
+| Training Loss | Epoch   | Step | Validation Loss |
+|:-------------:|:-------:|:----:|:---------------:|
+| 1.6578        | 1.3333  | 5    | 1.6863          |
+| 1.6277        | 2.6667  | 10   | 1.6771          |
+| 1.615         | 4.0     | 15   | 1.6615          |
+| 1.5879        | 5.3333  | 20   | 1.6372          |
+| 1.5835        | 6.6667  | 25   | 1.6028          |
+| 1.5908        | 8.0     | 30   | 1.5586          |
+| 1.5143        | 9.3333  | 35   | 1.5012          |
+| 1.4633        | 10.6667 | 40   | 1.4352          |
+| 1.3414        | 12.0    | 45   | 1.3606          |
+| 1.3229        | 13.3333 | 50   | 1.2811          |
+| 1.2218        | 14.6667 | 55   | 1.2119          |
+| 1.1352        | 16.0    | 60   | 1.1488          |
+| 1.0852        | 17.3333 | 65   | 1.0885          |
+| 0.9989        | 18.6667 | 70   | 1.0299          |
+| 0.9959        | 20.0    | 75   | 0.9757          |
+| 0.921         | 21.3333 | 80   | 0.9205          |
+| 0.8727        | 22.6667 | 85   | 0.8683          |
+| 0.8067        | 24.0    | 90   | 0.8200          |
+| 0.7785        | 25.3333 | 95   | 0.7783          |
+| 0.7139        | 26.6667 | 100  | 0.7396          |
+| 0.7081        | 28.0    | 105  | 0.7095          |
+| 0.6705        | 29.3333 | 110  | 0.6824          |
+| 0.6177        | 30.6667 | 115  | 0.6613          |
+| 0.6106        | 32.0    | 120  | 0.6418          |
+| 0.575         | 33.3333 | 125  | 0.6239          |
+| 0.5904        | 34.6667 | 130  | 0.6083          |
+| 0.5917        | 36.0    | 135  | 0.5927          |
+| 0.5051        | 37.3333 | 140  | 0.5801          |
+| 0.5169        | 38.6667 | 145  | 0.5656          |
+| 0.5442        | 40.0    | 150  | 0.5542          |
+| 0.5112        | 41.3333 | 155  | 0.5432          |
+| 0.5061        | 42.6667 | 160  | 0.5321          |
+| 0.5071        | 44.0    | 165  | 0.5234          |
+| 0.4373        | 45.3333 | 170  | 0.5119          |
+| 0.4476        | 46.6667 | 175  | 0.5049          |
+| 0.3914        | 48.0    | 180  | 0.4972          |
+| 0.465         | 49.3333 | 185  | 0.4914          |
+| 0.4122        | 50.6667 | 190  | 0.4890          |
+| 0.4209        | 52.0    | 195  | 0.4837          |
+| 0.3933        | 53.3333 | 200  | 0.4784          |
+| 0.3583        | 54.6667 | 205  | 0.4760          |
+| 0.3952        | 56.0    | 210  | 0.4727          |
+| 0.3858        | 57.3333 | 215  | 0.4708          |
+| 0.3433        | 58.6667 | 220  | 0.4707          |
+| 0.4041        | 60.0    | 225  | 0.4680          |
+| 0.3558        | 61.3333 | 230  | 0.4665          |
+| 0.382         | 62.6667 | 235  | 0.4650          |
+| 0.3625        | 64.0    | 240  | 0.4638          |
+| 0.3513        | 65.3333 | 245  | 0.4644          |
+| 0.3541        | 66.6667 | 250  | 0.4630          |
 
 
 ### Framework versions
 
-- PEFT 0.13.0
-- Transformers 4.45.1
-- Pytorch 2.4.1+cu121
+- PEFT 0.13.2
+- Transformers 4.45.2
+- Pytorch 2.4.0
 - Datasets 3.0.1
 - Tokenizers 0.20.0

adapter_config.json

@@ -1,7 +1,7 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
+  "base_model_name_or_path": "microsoft/Phi-3.5-mini-instruct",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,
@@ -10,7 +10,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 16,
+  "lora_alpha": 32,
   "lora_dropout": 0.05,
   "megatron_config": null,
   "megatron_core": "megatron.core",
@@ -20,13 +20,10 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "up_proj",
-    "o_proj",
-    "k_proj",
-    "gate_proj",
     "down_proj",
-    "q_proj"
+    "o_proj",
+    "gate_up_proj",
+    "qkv_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9683bef0e3f11bbdb5fef4040748cefcdfce22840d306c6861c340871b8d49d2
-size 35668592
+oid sha256:34328d550a8b27f1ce0567efebc0753762b2e56158e73fcf2fbfe056028991de
+size 100697728

tokenizer.json

@@ -2,7 +2,7 @@
   "version": "1.0",
   "truncation": {
     "direction": "Right",
-    "max_length": 512,
+    "max_length": 2048,
     "strategy": "LongestFirst",
     "stride": 0
   },

tokenizer_config.json

@@ -117,11 +117,11 @@
     }
   },
   "bos_token": "<s>",
-  "chat_template": "{% for message in messages %}{% if message['role'] == 'system' %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
+  "chat_template": "{% for message in messages %}{% if message['role'] == 'system' and message['content'] %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|endoftext|>",
   "legacy": false,
-  "model_max_length": 4096,
+  "model_max_length": 131072,
   "pad_token": "<unk>",
   "padding_side": "left",
   "sp_model_kwargs": {},

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0ce710ba59e50ddc28a07ae112ed2d16849ac7630a3c9206dbc26cb350177409
+oid sha256:3083cc2b7d42aa4e26b73ee560dd2185afd1e4c08af814236bcfcf5a7c7f3fb3
 size 5432