End of training
Browse files- README.md +15 -11
- all_results.json +16 -0
- config.json +40 -0
- eval_results.json +19 -0
- runs/Jul16_15-39-24_node29/events.out.tfevents.1721232190.node29.1097085.1 +3 -0
README.md
CHANGED
@@ -2,10 +2,14 @@
|
|
2 |
license: apache-2.0
|
3 |
library_name: peft
|
4 |
tags:
|
|
|
5 |
- trl
|
6 |
- dpo
|
7 |
- generated_from_trainer
|
8 |
base_model: openbmb/Eurus-7b-sft
|
|
|
|
|
|
|
9 |
model-index:
|
10 |
- name: eurus-dpo-qlora-uf-ours-uffull-5e-7
|
11 |
results: []
|
@@ -16,20 +20,20 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
|
17 |
# eurus-dpo-qlora-uf-ours-uffull-5e-7
|
18 |
|
19 |
-
This model is a fine-tuned version of [openbmb/Eurus-7b-sft](https://huggingface.co/openbmb/Eurus-7b-sft) on the
|
20 |
It achieves the following results on the evaluation set:
|
21 |
- Loss: 0.5784
|
22 |
-
- Rewards/chosen: -0.
|
23 |
-
- Rewards/rejected: -1.
|
24 |
-
- Rewards/accuracies: 0.
|
25 |
-
- Rewards/margins: 0.
|
26 |
-
- Rewards/margins Max: 2.
|
27 |
- Rewards/margins Min: -0.7893
|
28 |
-
- Rewards/margins Std: 1.
|
29 |
-
- Logps/rejected: -402.
|
30 |
-
- Logps/chosen: -353.
|
31 |
-
- Logits/rejected: -2.
|
32 |
-
- Logits/chosen: -2.
|
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: openbmb/Eurus-7b-sft
|
10 |
+
datasets:
|
11 |
+
- generation/UF
|
12 |
+
- generation/UFfull
|
13 |
model-index:
|
14 |
- name: eurus-dpo-qlora-uf-ours-uffull-5e-7
|
15 |
results: []
|
|
|
20 |
|
21 |
# eurus-dpo-qlora-uf-ours-uffull-5e-7
|
22 |
|
23 |
+
This model is a fine-tuned version of [openbmb/Eurus-7b-sft](https://huggingface.co/openbmb/Eurus-7b-sft) on the generation/UF and the generation/UFfull datasets.
|
24 |
It achieves the following results on the evaluation set:
|
25 |
- Loss: 0.5784
|
26 |
+
- Rewards/chosen: -0.8137
|
27 |
+
- Rewards/rejected: -1.4080
|
28 |
+
- Rewards/accuracies: 0.6890
|
29 |
+
- Rewards/margins: 0.5943
|
30 |
+
- Rewards/margins Max: 2.7059
|
31 |
- Rewards/margins Min: -0.7893
|
32 |
+
- Rewards/margins Std: 1.1535
|
33 |
+
- Logps/rejected: -402.0422
|
34 |
+
- Logps/chosen: -353.0313
|
35 |
+
- Logits/rejected: -2.0261
|
36 |
+
- Logits/chosen: -2.1316
|
37 |
|
38 |
## Model description
|
39 |
|
all_results.json
CHANGED
@@ -1,5 +1,21 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
"train_loss": 0.6004282865259383,
|
4 |
"train_runtime": 118629.0501,
|
5 |
"train_samples": 66812,
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"eval_logits/chosen": -2.13159441947937,
|
4 |
+
"eval_logits/rejected": -2.0261080265045166,
|
5 |
+
"eval_logps/chosen": -353.0312805175781,
|
6 |
+
"eval_logps/rejected": -402.0422058105469,
|
7 |
+
"eval_loss": 0.5784093141555786,
|
8 |
+
"eval_rewards/accuracies": 0.6890000104904175,
|
9 |
+
"eval_rewards/chosen": -0.8136823177337646,
|
10 |
+
"eval_rewards/margins": 0.5943145751953125,
|
11 |
+
"eval_rewards/margins_max": 2.7059268951416016,
|
12 |
+
"eval_rewards/margins_min": -0.7893158197402954,
|
13 |
+
"eval_rewards/margins_std": 1.1535371541976929,
|
14 |
+
"eval_rewards/rejected": -1.4079967737197876,
|
15 |
+
"eval_runtime": 1498.5961,
|
16 |
+
"eval_samples": 4000,
|
17 |
+
"eval_samples_per_second": 2.669,
|
18 |
+
"eval_steps_per_second": 0.167,
|
19 |
"train_loss": 0.6004282865259383,
|
20 |
"train_runtime": 118629.0501,
|
21 |
"train_samples": 66812,
|
config.json
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "openbmb/Eurus-7b-sft",
|
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": -2.13159441947937,
|
4 |
+
"eval_logits/rejected": -2.0261080265045166,
|
5 |
+
"eval_logps/chosen": -353.0312805175781,
|
6 |
+
"eval_logps/rejected": -402.0422058105469,
|
7 |
+
"eval_loss": 0.5784093141555786,
|
8 |
+
"eval_rewards/accuracies": 0.6890000104904175,
|
9 |
+
"eval_rewards/chosen": -0.8136823177337646,
|
10 |
+
"eval_rewards/margins": 0.5943145751953125,
|
11 |
+
"eval_rewards/margins_max": 2.7059268951416016,
|
12 |
+
"eval_rewards/margins_min": -0.7893158197402954,
|
13 |
+
"eval_rewards/margins_std": 1.1535371541976929,
|
14 |
+
"eval_rewards/rejected": -1.4079967737197876,
|
15 |
+
"eval_runtime": 1498.5961,
|
16 |
+
"eval_samples": 4000,
|
17 |
+
"eval_samples_per_second": 2.669,
|
18 |
+
"eval_steps_per_second": 0.167
|
19 |
+
}
|
runs/Jul16_15-39-24_node29/events.out.tfevents.1721232190.node29.1097085.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8c7aa2333d5c8c9b069bf71041d8f74e74b3a75e5fd227d1ca9a7ba655097011
|
3 |
+
size 1017
|