Commit
•
68283c8
1
Parent(s):
947f16b
End of training
Browse files
README.md
CHANGED
|
@@ -1,24 +1,23 @@
|
|
| 1 |
---
|
| 2 |
-
base_model: ahxt/
|
| 3 |
library_name: peft
|
| 4 |
-
license: mit
|
| 5 |
tags:
|
| 6 |
- trl
|
| 7 |
- sft
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
-
- name:
|
| 11 |
results: []
|
| 12 |
---
|
| 13 |
|
| 14 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
|
| 17 |
-
#
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [ahxt/
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss:
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
@@ -37,14 +36,14 @@ More information needed
|
|
| 37 |
### Training hyperparameters
|
| 38 |
|
| 39 |
The following hyperparameters were used during training:
|
| 40 |
-
- learning_rate: 0.
|
| 41 |
-
- train_batch_size:
|
| 42 |
- eval_batch_size: 32
|
| 43 |
- seed: 42
|
| 44 |
-
- gradient_accumulation_steps:
|
| 45 |
-
- total_train_batch_size:
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
-
- lr_scheduler_type:
|
| 48 |
- lr_scheduler_warmup_ratio: 0.03
|
| 49 |
- num_epochs: 10
|
| 50 |
|
|
@@ -52,18 +51,11 @@ The following hyperparameters were used during training:
|
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 54 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 55 |
-
|
|
| 56 |
-
|
|
| 57 |
-
|
|
| 58 |
-
|
|
| 59 |
-
|
|
| 60 |
-
| 2.1591 | 4.8 | 150 | 2.1316 |
|
| 61 |
-
| 2.1214 | 5.6 | 175 | 2.1011 |
|
| 62 |
-
| 2.0946 | 6.4 | 200 | 2.0781 |
|
| 63 |
-
| 2.0818 | 7.2 | 225 | 2.0622 |
|
| 64 |
-
| 2.0614 | 8.0 | 250 | 2.0528 |
|
| 65 |
-
| 2.0571 | 8.8 | 275 | 2.0485 |
|
| 66 |
-
| 2.0522 | 9.6 | 300 | 2.0471 |
|
| 67 |
|
| 68 |
|
| 69 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
+
base_model: ahxt/llama1_s_1.8B_experimental
|
| 3 |
library_name: peft
|
|
|
|
| 4 |
tags:
|
| 5 |
- trl
|
| 6 |
- sft
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
+
- name: Cold-Rec-LLama-2-7B
|
| 10 |
results: []
|
| 11 |
---
|
| 12 |
|
| 13 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
|
| 16 |
+
# Cold-Rec-LLama-2-7B
|
| 17 |
|
| 18 |
+
This model is a fine-tuned version of [ahxt/llama1_s_1.8B_experimental](https://huggingface.co/ahxt/llama1_s_1.8B_experimental) on the None dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.0777
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 36 |
### Training hyperparameters
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 0.001
|
| 40 |
+
- train_batch_size: 16
|
| 41 |
- eval_batch_size: 32
|
| 42 |
- seed: 42
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 64
|
| 45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: linear
|
| 47 |
- lr_scheduler_warmup_ratio: 0.03
|
| 48 |
- num_epochs: 10
|
| 49 |
|
|
|
|
| 51 |
|
| 52 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 53 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 54 |
+
| 0.2217 | 1.992 | 249 | 0.2119 |
|
| 55 |
+
| 0.1599 | 3.984 | 498 | 0.1558 |
|
| 56 |
+
| 0.1177 | 5.976 | 747 | 0.1157 |
|
| 57 |
+
| 0.0842 | 7.968 | 996 | 0.0875 |
|
| 58 |
+
| 0.075 | 9.96 | 1245 | 0.0776 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
|
| 61 |
### Framework versions
|
runs/Aug09_10-22-52_fastgpuserv/events.out.tfevents.1723268246.fastgpuserv.502349.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad26b338296efb151b9d75bb2d3adc053a8bd8b32894c14dd7fc8f83d4446db1
|
| 3 |
+
size 359
|