End of training
Browse files- README.md +14 -5
- adapter_config.json +4 -4
- adapter_model.safetensors +2 -2
- all_results.json +9 -5
- eval_results.json +5 -5
- metrics.json +1 -1
- train_results.json +5 -5
- trainer_state.json +228 -14
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -15,6 +15,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
# Mistral-7B-v0.1_case-briefs
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
|
|
|
|
|
|
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
@@ -33,19 +35,26 @@ More information needed
|
|
| 33 |
### Training hyperparameters
|
| 34 |
|
| 35 |
The following hyperparameters were used during training:
|
| 36 |
-
- learning_rate:
|
| 37 |
-
- train_batch_size:
|
| 38 |
-
- eval_batch_size:
|
| 39 |
- seed: 42
|
| 40 |
-
- gradient_accumulation_steps:
|
| 41 |
- total_train_batch_size: 16
|
| 42 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
- lr_scheduler_type: constant
|
| 44 |
- lr_scheduler_warmup_ratio: 0.03
|
| 45 |
-
-
|
| 46 |
|
| 47 |
### Training results
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
|
| 50 |
|
| 51 |
### Framework versions
|
|
|
|
| 15 |
# Mistral-7B-v0.1_case-briefs
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 1.1314
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
|
|
|
| 35 |
### Training hyperparameters
|
| 36 |
|
| 37 |
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 3e-05
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
- seed: 42
|
| 42 |
+
- gradient_accumulation_steps: 4
|
| 43 |
- total_train_batch_size: 16
|
| 44 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
- lr_scheduler_type: constant
|
| 46 |
- lr_scheduler_warmup_ratio: 0.03
|
| 47 |
+
- num_epochs: 2.0
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 52 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
| 53 |
+
| 1.1077 | 0.34 | 50 | 1.1506 |
|
| 54 |
+
| 1.0642 | 0.68 | 100 | 1.1310 |
|
| 55 |
+
| 1.0399 | 1.02 | 150 | 1.1266 |
|
| 56 |
+
| 1.0165 | 1.36 | 200 | 1.1302 |
|
| 57 |
+
| 1.0272 | 1.7 | 250 | 1.1314 |
|
| 58 |
|
| 59 |
|
| 60 |
### Framework versions
|
adapter_config.json
CHANGED
|
@@ -10,20 +10,20 @@
|
|
| 10 |
"layers_to_transform": null,
|
| 11 |
"loftq_config": {},
|
| 12 |
"lora_alpha": 16,
|
| 13 |
-
"lora_dropout": 0.
|
| 14 |
"megatron_config": null,
|
| 15 |
"megatron_core": "megatron.core",
|
| 16 |
"modules_to_save": null,
|
| 17 |
"peft_type": "LORA",
|
| 18 |
-
"r":
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
| 22 |
"down_proj",
|
| 23 |
-
"o_proj",
|
| 24 |
"gate_proj",
|
| 25 |
-
"q_proj",
|
| 26 |
"v_proj",
|
|
|
|
|
|
|
| 27 |
"up_proj",
|
| 28 |
"k_proj"
|
| 29 |
],
|
|
|
|
| 10 |
"layers_to_transform": null,
|
| 11 |
"loftq_config": {},
|
| 12 |
"lora_alpha": 16,
|
| 13 |
+
"lora_dropout": 0.1,
|
| 14 |
"megatron_config": null,
|
| 15 |
"megatron_core": "megatron.core",
|
| 16 |
"modules_to_save": null,
|
| 17 |
"peft_type": "LORA",
|
| 18 |
+
"r": 16,
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
| 22 |
"down_proj",
|
|
|
|
| 23 |
"gate_proj",
|
|
|
|
| 24 |
"v_proj",
|
| 25 |
+
"o_proj",
|
| 26 |
+
"q_proj",
|
| 27 |
"up_proj",
|
| 28 |
"k_proj"
|
| 29 |
],
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:110c0af297f717ff2cf7806e9ffaee869b95756862078bd045b2a2b00c46f728
|
| 3 |
+
size 83946192
|
all_results.json
CHANGED
|
@@ -1,7 +1,11 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 0
|
| 3 |
-
"
|
| 4 |
-
"
|
| 5 |
-
"
|
| 6 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_loss": 1.2288233041763306,
|
| 4 |
+
"eval_runtime": 22.0177,
|
| 5 |
+
"eval_samples_per_second": 11.899,
|
| 6 |
+
"eval_steps_per_second": 1.499,
|
| 7 |
+
"train_loss": 1.0460854338950851,
|
| 8 |
+
"train_runtime": 864.371,
|
| 9 |
+
"train_samples_per_second": 5.451,
|
| 10 |
+
"train_steps_per_second": 0.34
|
| 11 |
}
|
eval_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss": 1.
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second":
|
| 6 |
-
"eval_steps_per_second":
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 0.01,
|
| 3 |
+
"eval_loss": 1.2288233041763306,
|
| 4 |
+
"eval_runtime": 22.0177,
|
| 5 |
+
"eval_samples_per_second": 11.899,
|
| 6 |
+
"eval_steps_per_second": 1.499
|
| 7 |
}
|
metrics.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"run_name": "./output", "train_runtime":
|
|
|
|
| 1 |
+
{"run_name": "./output", "train_runtime": 13.1306, "train_samples_per_second": 1.219, "train_steps_per_second": 0.076, "train_loss": 1.3776695728302002, "epoch": 0.01, "eval_loss": 1.2288233041763306, "eval_runtime": 22.0177, "eval_samples_per_second": 11.899, "eval_steps_per_second": 1.499}
|
train_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 0
|
| 3 |
-
"train_loss": 1.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples_per_second":
|
| 6 |
-
"train_steps_per_second": 0.
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"train_loss": 1.0460854338950851,
|
| 4 |
+
"train_runtime": 864.371,
|
| 5 |
+
"train_samples_per_second": 5.451,
|
| 6 |
+
"train_steps_per_second": 0.34
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -1,30 +1,244 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"eval_steps":
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"
|
| 14 |
-
"
|
| 15 |
-
"
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
}
|
| 20 |
],
|
| 21 |
"logging_steps": 10,
|
| 22 |
-
"max_steps":
|
| 23 |
"num_input_tokens_seen": 0,
|
| 24 |
-
"num_train_epochs":
|
| 25 |
"save_steps": 250,
|
| 26 |
-
"total_flos":
|
| 27 |
-
"train_batch_size":
|
| 28 |
"trial_name": null,
|
| 29 |
"trial_params": null
|
| 30 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.99660441426146,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 294,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.07,
|
| 13 |
+
"learning_rate": 3e-05,
|
| 14 |
+
"loss": 1.2065,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.14,
|
| 19 |
+
"learning_rate": 3e-05,
|
| 20 |
+
"loss": 1.1456,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.2,
|
| 25 |
+
"learning_rate": 3e-05,
|
| 26 |
+
"loss": 1.1411,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.27,
|
| 31 |
+
"learning_rate": 3e-05,
|
| 32 |
+
"loss": 1.1179,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.34,
|
| 37 |
+
"learning_rate": 3e-05,
|
| 38 |
+
"loss": 1.1077,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.34,
|
| 43 |
+
"eval_loss": 1.1505802869796753,
|
| 44 |
+
"eval_runtime": 15.5075,
|
| 45 |
+
"eval_samples_per_second": 16.895,
|
| 46 |
+
"eval_steps_per_second": 4.256,
|
| 47 |
+
"step": 50
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.41,
|
| 51 |
+
"learning_rate": 3e-05,
|
| 52 |
+
"loss": 1.0964,
|
| 53 |
+
"step": 60
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.48,
|
| 57 |
+
"learning_rate": 3e-05,
|
| 58 |
+
"loss": 1.0923,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.54,
|
| 63 |
+
"learning_rate": 3e-05,
|
| 64 |
+
"loss": 1.1404,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.61,
|
| 69 |
+
"learning_rate": 3e-05,
|
| 70 |
+
"loss": 1.1145,
|
| 71 |
+
"step": 90
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.68,
|
| 75 |
+
"learning_rate": 3e-05,
|
| 76 |
+
"loss": 1.0642,
|
| 77 |
+
"step": 100
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.68,
|
| 81 |
+
"eval_loss": 1.1310365200042725,
|
| 82 |
+
"eval_runtime": 15.5192,
|
| 83 |
+
"eval_samples_per_second": 16.882,
|
| 84 |
+
"eval_steps_per_second": 4.253,
|
| 85 |
+
"step": 100
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.75,
|
| 89 |
+
"learning_rate": 3e-05,
|
| 90 |
+
"loss": 1.0558,
|
| 91 |
+
"step": 110
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 0.81,
|
| 95 |
+
"learning_rate": 3e-05,
|
| 96 |
+
"loss": 1.1053,
|
| 97 |
+
"step": 120
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.88,
|
| 101 |
+
"learning_rate": 3e-05,
|
| 102 |
+
"loss": 1.0948,
|
| 103 |
+
"step": 130
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 0.95,
|
| 107 |
+
"learning_rate": 3e-05,
|
| 108 |
+
"loss": 1.0725,
|
| 109 |
+
"step": 140
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 1.02,
|
| 113 |
+
"learning_rate": 3e-05,
|
| 114 |
+
"loss": 1.0399,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 1.02,
|
| 119 |
+
"eval_loss": 1.1266452074050903,
|
| 120 |
+
"eval_runtime": 15.5722,
|
| 121 |
+
"eval_samples_per_second": 16.825,
|
| 122 |
+
"eval_steps_per_second": 4.238,
|
| 123 |
+
"step": 150
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.09,
|
| 127 |
+
"learning_rate": 3e-05,
|
| 128 |
+
"loss": 0.9873,
|
| 129 |
+
"step": 160
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.15,
|
| 133 |
+
"learning_rate": 3e-05,
|
| 134 |
+
"loss": 1.0614,
|
| 135 |
+
"step": 170
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 1.22,
|
| 139 |
+
"learning_rate": 3e-05,
|
| 140 |
+
"loss": 0.979,
|
| 141 |
+
"step": 180
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 1.29,
|
| 145 |
+
"learning_rate": 3e-05,
|
| 146 |
+
"loss": 0.9585,
|
| 147 |
+
"step": 190
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.36,
|
| 151 |
+
"learning_rate": 3e-05,
|
| 152 |
+
"loss": 1.0165,
|
| 153 |
+
"step": 200
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.36,
|
| 157 |
+
"eval_loss": 1.1301801204681396,
|
| 158 |
+
"eval_runtime": 15.5114,
|
| 159 |
+
"eval_samples_per_second": 16.891,
|
| 160 |
+
"eval_steps_per_second": 4.255,
|
| 161 |
+
"step": 200
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.43,
|
| 165 |
+
"learning_rate": 3e-05,
|
| 166 |
+
"loss": 0.987,
|
| 167 |
+
"step": 210
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.49,
|
| 171 |
+
"learning_rate": 3e-05,
|
| 172 |
+
"loss": 0.9076,
|
| 173 |
+
"step": 220
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.56,
|
| 177 |
+
"learning_rate": 3e-05,
|
| 178 |
+
"loss": 0.9922,
|
| 179 |
+
"step": 230
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.63,
|
| 183 |
+
"learning_rate": 3e-05,
|
| 184 |
+
"loss": 0.9889,
|
| 185 |
+
"step": 240
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.7,
|
| 189 |
+
"learning_rate": 3e-05,
|
| 190 |
+
"loss": 1.0272,
|
| 191 |
+
"step": 250
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.7,
|
| 195 |
+
"eval_loss": 1.1314338445663452,
|
| 196 |
+
"eval_runtime": 15.5119,
|
| 197 |
+
"eval_samples_per_second": 16.89,
|
| 198 |
+
"eval_steps_per_second": 4.255,
|
| 199 |
+
"step": 250
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 1.77,
|
| 203 |
+
"learning_rate": 3e-05,
|
| 204 |
+
"loss": 0.977,
|
| 205 |
+
"step": 260
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 1.83,
|
| 209 |
+
"learning_rate": 3e-05,
|
| 210 |
+
"loss": 1.0425,
|
| 211 |
+
"step": 270
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 1.9,
|
| 215 |
+
"learning_rate": 3e-05,
|
| 216 |
+
"loss": 0.9611,
|
| 217 |
+
"step": 280
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.97,
|
| 221 |
+
"learning_rate": 3e-05,
|
| 222 |
+
"loss": 0.9305,
|
| 223 |
+
"step": 290
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 2.0,
|
| 227 |
+
"step": 294,
|
| 228 |
+
"total_flos": 6.900509298543821e+16,
|
| 229 |
+
"train_loss": 1.0460854338950851,
|
| 230 |
+
"train_runtime": 864.371,
|
| 231 |
+
"train_samples_per_second": 5.451,
|
| 232 |
+
"train_steps_per_second": 0.34
|
| 233 |
}
|
| 234 |
],
|
| 235 |
"logging_steps": 10,
|
| 236 |
+
"max_steps": 294,
|
| 237 |
"num_input_tokens_seen": 0,
|
| 238 |
+
"num_train_epochs": 2,
|
| 239 |
"save_steps": 250,
|
| 240 |
+
"total_flos": 6.900509298543821e+16,
|
| 241 |
+
"train_batch_size": 4,
|
| 242 |
"trial_name": null,
|
| 243 |
"trial_params": null
|
| 244 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6648
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:670309ed11279de039411d6ea6d7cb68f8f00d95a334c639a77e271f6292bb0f
|
| 3 |
size 6648
|