End of training
Browse files- README.md +16 -13
- all_results.json +16 -0
- config.json +40 -0
- eval_results.json +19 -0
- runs/Jul24_19-27-40_node03/events.out.tfevents.1721823719.node03.3774365.1 +3 -0
README.md
CHANGED
@@ -2,10 +2,13 @@
|
|
2 |
license: apache-2.0
|
3 |
library_name: peft
|
4 |
tags:
|
|
|
5 |
- trl
|
6 |
- dpo
|
7 |
- generated_from_trainer
|
8 |
base_model: alignment-handbook/zephyr-7b-sft-full
|
|
|
|
|
9 |
model-index:
|
10 |
- name: zephyr-7b-dpo-oursuf6k-qlora-5e-6
|
11 |
results: []
|
@@ -16,20 +19,20 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
|
17 |
# zephyr-7b-dpo-oursuf6k-qlora-5e-6
|
18 |
|
19 |
-
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the
|
20 |
It achieves the following results on the evaluation set:
|
21 |
-
- Loss: 0.
|
22 |
-
- Rewards/chosen: -2.
|
23 |
-
- Rewards/rejected: -2.
|
24 |
-
- Rewards/accuracies: 0.
|
25 |
-
- Rewards/margins: 0.
|
26 |
-
- Rewards/margins Max: 2.
|
27 |
-
- Rewards/margins Min: -0.
|
28 |
-
- Rewards/margins Std: 1.
|
29 |
-
- Logps/rejected: -
|
30 |
-
- Logps/chosen: -
|
31 |
-
- Logits/rejected: -1.
|
32 |
-
- Logits/chosen: -1.
|
33 |
|
34 |
## Model description
|
35 |
|
|
|
2 |
license: apache-2.0
|
3 |
library_name: peft
|
4 |
tags:
|
5 |
+
- alignment-handbook
|
6 |
- trl
|
7 |
- dpo
|
8 |
- generated_from_trainer
|
9 |
base_model: alignment-handbook/zephyr-7b-sft-full
|
10 |
+
datasets:
|
11 |
+
- generation/UF6k
|
12 |
model-index:
|
13 |
- name: zephyr-7b-dpo-oursuf6k-qlora-5e-6
|
14 |
results: []
|
|
|
19 |
|
20 |
# zephyr-7b-dpo-oursuf6k-qlora-5e-6
|
21 |
|
22 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the generation/UF6k dataset.
|
23 |
It achieves the following results on the evaluation set:
|
24 |
+
- Loss: 0.5972
|
25 |
+
- Rewards/chosen: -2.1562
|
26 |
+
- Rewards/rejected: -2.9617
|
27 |
+
- Rewards/accuracies: 0.6865
|
28 |
+
- Rewards/margins: 0.8055
|
29 |
+
- Rewards/margins Max: 2.4253
|
30 |
+
- Rewards/margins Min: -0.7592
|
31 |
+
- Rewards/margins Std: 1.4237
|
32 |
+
- Logps/rejected: -555.3531
|
33 |
+
- Logps/chosen: -500.8411
|
34 |
+
- Logits/rejected: -1.8009
|
35 |
+
- Logits/chosen: -1.8449
|
36 |
|
37 |
## Model description
|
38 |
|
all_results.json
CHANGED
@@ -1,5 +1,21 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
"train_loss": 0.42957196057888497,
|
4 |
"train_runtime": 6346.2002,
|
5 |
"train_samples": 10941,
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"eval_logits/chosen": -1.844919204711914,
|
4 |
+
"eval_logits/rejected": -1.8009027242660522,
|
5 |
+
"eval_logps/chosen": -500.84112548828125,
|
6 |
+
"eval_logps/rejected": -555.3530883789062,
|
7 |
+
"eval_loss": 0.5972276329994202,
|
8 |
+
"eval_rewards/accuracies": 0.6865079402923584,
|
9 |
+
"eval_rewards/chosen": -2.1561989784240723,
|
10 |
+
"eval_rewards/margins": 0.8055059313774109,
|
11 |
+
"eval_rewards/margins_max": 2.425264835357666,
|
12 |
+
"eval_rewards/margins_min": -0.7591551542282104,
|
13 |
+
"eval_rewards/margins_std": 1.4236739873886108,
|
14 |
+
"eval_rewards/rejected": -2.961704730987549,
|
15 |
+
"eval_runtime": 282.7661,
|
16 |
+
"eval_samples": 2000,
|
17 |
+
"eval_samples_per_second": 7.073,
|
18 |
+
"eval_steps_per_second": 0.223,
|
19 |
"train_loss": 0.42957196057888497,
|
20 |
"train_runtime": 6346.2002,
|
21 |
"train_samples": 10941,
|
config.json
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "alignment-handbook/zephyr-7b-sft-full",
|
3 |
+
"architectures": [
|
4 |
+
"MistralForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 1,
|
8 |
+
"eos_token_id": 2,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 4096,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"intermediate_size": 14336,
|
13 |
+
"max_position_embeddings": 32768,
|
14 |
+
"model_type": "mistral",
|
15 |
+
"num_attention_heads": 32,
|
16 |
+
"num_hidden_layers": 32,
|
17 |
+
"num_key_value_heads": 8,
|
18 |
+
"quantization_config": {
|
19 |
+
"_load_in_4bit": true,
|
20 |
+
"_load_in_8bit": false,
|
21 |
+
"bnb_4bit_compute_dtype": "bfloat16",
|
22 |
+
"bnb_4bit_quant_type": "nf4",
|
23 |
+
"bnb_4bit_use_double_quant": false,
|
24 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
25 |
+
"llm_int8_has_fp16_weight": false,
|
26 |
+
"llm_int8_skip_modules": null,
|
27 |
+
"llm_int8_threshold": 6.0,
|
28 |
+
"load_in_4bit": true,
|
29 |
+
"load_in_8bit": false,
|
30 |
+
"quant_method": "bitsandbytes"
|
31 |
+
},
|
32 |
+
"rms_norm_eps": 1e-05,
|
33 |
+
"rope_theta": 10000.0,
|
34 |
+
"sliding_window": 4096,
|
35 |
+
"tie_word_embeddings": false,
|
36 |
+
"torch_dtype": "bfloat16",
|
37 |
+
"transformers_version": "4.39.0.dev0",
|
38 |
+
"use_cache": true,
|
39 |
+
"vocab_size": 32000
|
40 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"eval_logits/chosen": -1.844919204711914,
|
4 |
+
"eval_logits/rejected": -1.8009027242660522,
|
5 |
+
"eval_logps/chosen": -500.84112548828125,
|
6 |
+
"eval_logps/rejected": -555.3530883789062,
|
7 |
+
"eval_loss": 0.5972276329994202,
|
8 |
+
"eval_rewards/accuracies": 0.6865079402923584,
|
9 |
+
"eval_rewards/chosen": -2.1561989784240723,
|
10 |
+
"eval_rewards/margins": 0.8055059313774109,
|
11 |
+
"eval_rewards/margins_max": 2.425264835357666,
|
12 |
+
"eval_rewards/margins_min": -0.7591551542282104,
|
13 |
+
"eval_rewards/margins_std": 1.4236739873886108,
|
14 |
+
"eval_rewards/rejected": -2.961704730987549,
|
15 |
+
"eval_runtime": 282.7661,
|
16 |
+
"eval_samples": 2000,
|
17 |
+
"eval_samples_per_second": 7.073,
|
18 |
+
"eval_steps_per_second": 0.223
|
19 |
+
}
|
runs/Jul24_19-27-40_node03/events.out.tfevents.1721823719.node03.3774365.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a471cae7d2e87e18447c211ce05ae0d9d9aaae763beeaada5aa441243f73bf7
|
3 |
+
size 1017
|