Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/train.log +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/model.safetensors +3 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/tokenizer.json +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/tokenizer_config.json +13 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/train.log +0 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_word_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch14_step20020_loss9.1443_rougel24.5158/config.json +42 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_word_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch14_step20020_loss9.1443_rougel24.5158/generation_config.json +6 -0
- gpt2/gpt2-base/mta_dskd_v2_ablation_word_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch14_step20020_loss9.1443_rougel24.5158/model.safetensors +3 -0
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac189debeac50c85f66bc18400d06dcd576e29c13a78006ba13d4aa453370802
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3533_rougel24.9879/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32687ca4e0b7387c45916d30b8453f35c27c4631b4d514f30356b3821b4d4411
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4173_rougel24.8735/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d1429fce6fe0ab16baeb9a6222f1b7058d60b88116d3b36dbc9d63c98aa03fc
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4356_rougel24.8686/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b3817c9ef0cc67af3827612f35a36200e072637c562aa314d1906ed8bd1d6462
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4688_rougel25.0958/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c82df7529d614a9cfe705ece4a6050250a1b5b3128d2a252de301603f696f4b1
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch20_step28600_loss9.4765_rougel24.9160/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_phrase_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/train.log
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c154b44a396ab868945047abe21a84eedfdb066116055e06b5207fc6e464ecc9
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch16_step22880_loss9.3359_rougel24.7419/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:de97d538bb59ae8d479bec3691f85cd8b63c366fa4802a03799080d93abd3806
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch17_step24310_loss9.4317_rougel24.6494/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd7d63ccdc647ac5ee15b58d80a10852c9d0c94d3515aed3c621b49f14241cf1
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch18_step25740_loss9.4385_rougel24.5294/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a9f3cfa3dae5d456f6a3b7913e108ee15d23f3056e63ae8a31433f4ed26e21e
|
| 3 |
+
size 248894656
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch19_step27170_loss9.4712_rougel24.7677/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": true,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_wo_weight/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/train.log
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2/gpt2-base/mta_dskd_v2_ablation_word_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch14_step20020_loss9.1443_rougel24.5158/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"is_model_parallel": false,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"model_type": "gpt2",
|
| 16 |
+
"n_ctx": 1024,
|
| 17 |
+
"n_embd": 768,
|
| 18 |
+
"n_head": 12,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 12,
|
| 21 |
+
"n_positions": 1024,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"reorder_and_upcast_attn": false,
|
| 24 |
+
"resid_pdrop": 0.1,
|
| 25 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 26 |
+
"scale_attn_weights": true,
|
| 27 |
+
"summary_activation": null,
|
| 28 |
+
"summary_first_dropout": 0.1,
|
| 29 |
+
"summary_proj_to_labels": true,
|
| 30 |
+
"summary_type": "cls_index",
|
| 31 |
+
"summary_use_proj": true,
|
| 32 |
+
"task_specific_params": {
|
| 33 |
+
"text-generation": {
|
| 34 |
+
"do_sample": true,
|
| 35 |
+
"max_length": 50
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"tie_word_embeddings": true,
|
| 39 |
+
"transformers_version": "5.9.0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 50257
|
| 42 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_word_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch14_step20020_loss9.1443_rougel24.5158/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.9.0"
|
| 6 |
+
}
|
gpt2/gpt2-base/mta_dskd_v2_ablation_word_level/forward_kl-bf16__teacher_qwen1.5__kd^rate0.5__kd^temp2.0__epoch20__bsz8x1x1x1_8__lr0.0005__proj^lr0.001/epoch14_step20020_loss9.1443_rougel24.5158/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ce2f2281077a617482f7889002c552466d837f7471edae043ba66bdc0b21635
|
| 3 |
+
size 248894656
|