Upload folder using huggingface_hub
Browse files- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/README.md +9 -0
- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/adapter_config.json +26 -0
- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/adapter_model.bin +3 -0
- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/config.json +68 -0
- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/generation_config.json +14 -0
- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/non_lora_trainables.bin +3 -0
- llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/trainer_state.json +0 -0
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/adapter_config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "/data/wiedmann/hub/models--lmms-lab--llava-onevision-qwen2-7b-ov/snapshots/0b07bf7565e244cf4f39982249eafe8cd799d6dd",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 32,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 64,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"gate_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"o_proj",
|
| 20 |
+
"v_proj",
|
| 21 |
+
"up_proj",
|
| 22 |
+
"k_proj",
|
| 23 |
+
"down_proj"
|
| 24 |
+
],
|
| 25 |
+
"task_type": "CAUSAL_LM"
|
| 26 |
+
}
|
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dae8934b14f78f93271b7d4f86516e3c0fa37acd2e77fae2f298e6ec2bc5d882
|
| 3 |
+
size 346162458
|
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/config.json
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/data/wiedmann/hub/models--lmms-lab--llava-onevision-qwen2-7b-ov/snapshots/0b07bf7565e244cf4f39982249eafe8cd799d6dd",
|
| 3 |
+
"add_faster_video": false,
|
| 4 |
+
"add_time_instruction": false,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlavaQwenForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"bos_token_id": 151643,
|
| 10 |
+
"eos_token_id": 151645,
|
| 11 |
+
"faster_token_stride": 10,
|
| 12 |
+
"force_sample": false,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 3584,
|
| 15 |
+
"image_aspect_ratio": "square",
|
| 16 |
+
"image_crop_resolution": null,
|
| 17 |
+
"image_grid_pinpoints": null,
|
| 18 |
+
"image_split_resolution": null,
|
| 19 |
+
"image_token_index": 151646,
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 18944,
|
| 22 |
+
"max_position_embeddings": 32768,
|
| 23 |
+
"max_window_layers": 28,
|
| 24 |
+
"mm_hidden_size": 1152,
|
| 25 |
+
"mm_newline_position": "no_token",
|
| 26 |
+
"mm_patch_merge_type": "spatial_unpad",
|
| 27 |
+
"mm_projector_lr": null,
|
| 28 |
+
"mm_projector_type": "mlp2x_gelu",
|
| 29 |
+
"mm_resampler_type": "streaming_agg",
|
| 30 |
+
"mm_spatial_pool_mode": "bilinear",
|
| 31 |
+
"mm_spatial_pool_stride": null,
|
| 32 |
+
"mm_streaming_frames_per_chunk": 4,
|
| 33 |
+
"mm_streaming_input_dim": 1152,
|
| 34 |
+
"mm_streaming_mlp_ratio": 4.0,
|
| 35 |
+
"mm_streaming_num_heads": 8,
|
| 36 |
+
"mm_streaming_num_layers": 4,
|
| 37 |
+
"mm_streaming_num_state_tokens": 2048,
|
| 38 |
+
"mm_streaming_patches_per_frame": 729,
|
| 39 |
+
"mm_streaming_state_dim": 1152,
|
| 40 |
+
"mm_streaming_vision_chunk_size": 8,
|
| 41 |
+
"mm_tunable_parts": "mm_mlp_adapter,mm_vision_resampler",
|
| 42 |
+
"mm_use_im_patch_token": false,
|
| 43 |
+
"mm_use_im_start_end": false,
|
| 44 |
+
"mm_vision_select_feature": "patch",
|
| 45 |
+
"mm_vision_select_layer": -2,
|
| 46 |
+
"mm_vision_tower": "google/siglip-so400m-patch14-384",
|
| 47 |
+
"mm_vision_tower_lr": null,
|
| 48 |
+
"model_type": "llava_qwen",
|
| 49 |
+
"num_attention_heads": 28,
|
| 50 |
+
"num_hidden_layers": 28,
|
| 51 |
+
"num_key_value_heads": 4,
|
| 52 |
+
"pos_skipping_range": 4096,
|
| 53 |
+
"rms_norm_eps": 1e-06,
|
| 54 |
+
"rope_scaling": null,
|
| 55 |
+
"rope_theta": 1000000.0,
|
| 56 |
+
"sliding_window": 131072,
|
| 57 |
+
"tie_word_embeddings": false,
|
| 58 |
+
"tokenizer_model_max_length": 8192,
|
| 59 |
+
"tokenizer_padding_side": "right",
|
| 60 |
+
"torch_dtype": "bfloat16",
|
| 61 |
+
"transformers_version": "4.40.0.dev0",
|
| 62 |
+
"use_cache": true,
|
| 63 |
+
"use_mm_proj": true,
|
| 64 |
+
"use_pos_skipping": false,
|
| 65 |
+
"use_sliding_window": false,
|
| 66 |
+
"vision_tower_pretrained": null,
|
| 67 |
+
"vocab_size": 152064
|
| 68 |
+
}
|
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151645,
|
| 6 |
+
151643
|
| 7 |
+
],
|
| 8 |
+
"pad_token_id": 151643,
|
| 9 |
+
"repetition_penalty": 1.05,
|
| 10 |
+
"temperature": 0.7,
|
| 11 |
+
"top_k": 20,
|
| 12 |
+
"top_p": 0.8,
|
| 13 |
+
"transformers_version": "4.40.0.dev0"
|
| 14 |
+
}
|
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/non_lora_trainables.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83df14521784f0469cb8816cfa0481571642869eba79790c45ec99df93989734
|
| 3 |
+
size 211441788
|
llavanext-google_siglip-so400m-patch14-384-Qwen_Qwen2-7B-Instruct-nextqa_shuffled-streaming_baseline_lora_corr_agg_fc4_s2048/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|