Commit
•
53f7e4d
1
Parent(s):
14c73cf
End of training
Browse files
README.md
CHANGED
|
@@ -1,10 +1,11 @@
|
|
| 1 |
---
|
| 2 |
base_model: facebook/opt-350m
|
| 3 |
datasets:
|
| 4 |
-
-
|
| 5 |
library_name: peft
|
| 6 |
license: other
|
| 7 |
tags:
|
|
|
|
| 8 |
- trl
|
| 9 |
- sft
|
| 10 |
- generated_from_trainer
|
|
@@ -18,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
# opt350m-qlora
|
| 20 |
|
| 21 |
-
This model is a fine-tuned version of [facebook/opt-350m](https://huggingface.co/facebook/opt-350m) on the
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
- Loss: 1.7868
|
| 24 |
|
|
|
|
| 1 |
---
|
| 2 |
base_model: facebook/opt-350m
|
| 3 |
datasets:
|
| 4 |
+
- HuggingFaceH4/ultrachat_200k
|
| 5 |
library_name: peft
|
| 6 |
license: other
|
| 7 |
tags:
|
| 8 |
+
- alignment-handbook
|
| 9 |
- trl
|
| 10 |
- sft
|
| 11 |
- generated_from_trainer
|
|
|
|
| 19 |
|
| 20 |
# opt350m-qlora
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/opt-350m](https://huggingface.co/facebook/opt-350m) on the HuggingFaceH4/ultrachat_200k dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
- Loss: 1.7868
|
| 25 |
|
all_results.json
CHANGED
|
@@ -1,5 +1,10 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9999380306128772,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
"total_flos": 4248917998829568.0,
|
| 4 |
"train_loss": 1.863956661347298,
|
| 5 |
"train_runtime": 6909.389,
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9999380306128772,
|
| 3 |
+
"eval_loss": 1.7868006229400635,
|
| 4 |
+
"eval_runtime": 207.2025,
|
| 5 |
+
"eval_samples": 23109,
|
| 6 |
+
"eval_samples_per_second": 68.913,
|
| 7 |
+
"eval_steps_per_second": 4.31,
|
| 8 |
"total_flos": 4248917998829568.0,
|
| 9 |
"train_loss": 1.863956661347298,
|
| 10 |
"train_runtime": 6909.389,
|
config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/opt-350m",
|
| 3 |
+
"_remove_final_layer_norm": false,
|
| 4 |
+
"activation_dropout": 0.0,
|
| 5 |
+
"activation_function": "relu",
|
| 6 |
+
"architectures": [
|
| 7 |
+
"OPTForCausalLM"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"bos_token_id": 2,
|
| 11 |
+
"do_layer_norm_before": false,
|
| 12 |
+
"dropout": 0.1,
|
| 13 |
+
"enable_bias": true,
|
| 14 |
+
"eos_token_id": 2,
|
| 15 |
+
"ffn_dim": 4096,
|
| 16 |
+
"hidden_size": 1024,
|
| 17 |
+
"init_std": 0.02,
|
| 18 |
+
"layer_norm_elementwise_affine": true,
|
| 19 |
+
"layerdrop": 0.0,
|
| 20 |
+
"max_position_embeddings": 2048,
|
| 21 |
+
"model_type": "opt",
|
| 22 |
+
"num_attention_heads": 16,
|
| 23 |
+
"num_hidden_layers": 24,
|
| 24 |
+
"pad_token_id": 1,
|
| 25 |
+
"prefix": "</s>",
|
| 26 |
+
"torch_dtype": "bfloat16",
|
| 27 |
+
"transformers_version": "4.45.0",
|
| 28 |
+
"use_cache": true,
|
| 29 |
+
"vocab_size": 50272,
|
| 30 |
+
"word_embed_proj_dim": 512
|
| 31 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9999380306128772,
|
| 3 |
+
"eval_loss": 1.7868006229400635,
|
| 4 |
+
"eval_runtime": 207.2025,
|
| 5 |
+
"eval_samples": 23109,
|
| 6 |
+
"eval_samples_per_second": 68.913,
|
| 7 |
+
"eval_steps_per_second": 4.31
|
| 8 |
+
}
|
runs/Nov06_12-39-19_node04.cluster/events.out.tfevents.1730932686.node04.cluster.3914975.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ffd4529ee2bbeed8d738c72fecbd4e19e19f14b836b93c469e10e99559b5580
|
| 3 |
+
size 359
|