Hugging Face's logo

glide-the
/
Chief_chain_generator-Qwen2_5-7B-Instruct-GPTQ-4bits

Safetensors
qwen2
4-bit precision
gptq
Model card Files
xet
Community
dmeck commited on
Commit
1ae3779
·
verified ·
1 Parent(s): 7bf48ae

Upload folder using huggingface_hub

Browse files
Files changed (17) hide show
  1. .gitattributes +1 -0
  2. added_tokens.json +24 -0
  3. chat_template.jinja +54 -0
  4. config.json +79 -0
  5. generation_config.json +14 -0
  6. glide-the/Chief_chain_generator-Qwen2.5-7B-Instruct-GPTQ-4bits-eora_rank64_c4/adapter_config.json +24 -0
  7. glide-the/Chief_chain_generator-Qwen2.5-7B-Instruct-GPTQ-4bits-eora_rank64_c4/adapter_model.safetensors +3 -0
  8. merges.txt +0 -0
  9. model-00001-of-00002.safetensors +3 -0
  10. model-00002-of-00002.safetensors +3 -0
  11. model.safetensors.index.json +934 -0
  12. quant_log.csv +197 -0
  13. quantize_config.json +21 -0
  14. special_tokens_map.json +25 -0
  15. tokenizer.json +3 -0
  16. tokenizer_config.json +208 -0
  17. vocab.json +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
chat_template.jinja ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0]['role'] == 'system' %}
4
+ {{- messages[0]['content'] }}
5
+ {%- else %}
6
+ {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
7
+ {%- endif %}
8
+ {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
9
+ {%- for tool in tools %}
10
+ {{- "\n" }}
11
+ {{- tool | tojson }}
12
+ {%- endfor %}
13
+ {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
14
+ {%- else %}
15
+ {%- if messages[0]['role'] == 'system' %}
16
+ {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
17
+ {%- else %}
18
+ {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
19
+ {%- endif %}
20
+ {%- endif %}
21
+ {%- for message in messages %}
22
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
23
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
24
+ {%- elif message.role == "assistant" %}
25
+ {{- '<|im_start|>' + message.role }}
26
+ {%- if message.content %}
27
+ {{- '\n' + message.content }}
28
+ {%- endif %}
29
+ {%- for tool_call in message.tool_calls %}
30
+ {%- if tool_call.function is defined %}
31
+ {%- set tool_call = tool_call.function %}
32
+ {%- endif %}
33
+ {{- '\n<tool_call>\n{"name": "' }}
34
+ {{- tool_call.name }}
35
+ {{- '", "arguments": ' }}
36
+ {{- tool_call.arguments | tojson }}
37
+ {{- '}\n</tool_call>' }}
38
+ {%- endfor %}
39
+ {{- '<|im_end|>\n' }}
40
+ {%- elif message.role == "tool" %}
41
+ {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
42
+ {{- '<|im_start|>user' }}
43
+ {%- endif %}
44
+ {{- '\n<tool_response>\n' }}
45
+ {{- message.content }}
46
+ {{- '\n</tool_response>' }}
47
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
48
+ {{- '<|im_end|>\n' }}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- endfor %}
52
+ {%- if add_generation_prompt %}
53
+ {{- '<|im_start|>assistant\n' }}
54
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151645,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 3584,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 18944,
12
+ "layer_types": [
13
+ "full_attention",
14
+ "full_attention",
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention"
41
+ ],
42
+ "max_position_embeddings": 32768,
43
+ "max_window_layers": 28,
44
+ "model_type": "qwen2",
45
+ "num_attention_heads": 28,
46
+ "num_hidden_layers": 28,
47
+ "num_key_value_heads": 4,
48
+ "quantization_config": {
49
+ "bits": 4,
50
+ "checkpoint_format": "gptq",
51
+ "desc_act": true,
52
+ "group_size": 128,
53
+ "lm_head": false,
54
+ "meta": {
55
+ "damp_auto_increment": 0.0025,
56
+ "damp_percent": 0.01,
57
+ "mse": 0.0,
58
+ "quantizer": [
59
+ "gptqmodel:2.2.0"
60
+ ],
61
+ "static_groups": false,
62
+ "true_sequential": true,
63
+ "uri": "https://github.com/modelcloud/gptqmodel"
64
+ },
65
+ "pack_dtype": "int32",
66
+ "quant_method": "gptq",
67
+ "sym": true
68
+ },
69
+ "rms_norm_eps": 1e-06,
70
+ "rope_scaling": null,
71
+ "rope_theta": 1000000.0,
72
+ "sliding_window": null,
73
+ "tie_word_embeddings": false,
74
+ "torch_dtype": "float16",
75
+ "transformers_version": "4.53.2",
76
+ "use_cache": true,
77
+ "use_sliding_window": false,
78
+ "vocab_size": 152064
79
+ }
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": "4.53.2"
14
+ }
glide-the/Chief_chain_generator-Qwen2.5-7B-Instruct-GPTQ-4bits-eora_rank64_c4/adapter_config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "glide-the/Chief_chain_generator-Qwen2.5-7B-Instruct-GPTQ-4bits",
3
+ "r": 64,
4
+ "target_modules": [
5
+ "v_proj",
6
+ "o_proj",
7
+ "gate_proj",
8
+ "up_proj",
9
+ "q_proj",
10
+ "down_proj",
11
+ "k_proj"
12
+ ],
13
+ "lora_alpha": 64,
14
+ "lora_dropout": 0.0,
15
+ "fan_in_fan_out": false,
16
+ "bias": "none",
17
+ "use_rslora": false,
18
+ "init_lora_weights": true,
19
+ "megatron_core": "megatron.core",
20
+ "use_dora": false,
21
+ "lora_bias": false,
22
+ "task_type": "CAUSAL_LM",
23
+ "peft_type": "LORA"
24
+ }
glide-the/Chief_chain_generator-Qwen2.5-7B-Instruct-GPTQ-4bits-eora_rank64_c4/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:782aaa2583306a8c558170f480bc28e26a1366804bcbd7ce4d1087139bfcfbc2
3
+ size 323014168
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:506df221f04afdf2c2f73e3feae273d3b8fbad86c5546a0f9449bb9c1ea4a835
3
+ size 3999172768
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec6abc3e6779ac599c40e67894ce5c1dc37d474c461f737da07bc1d495ff9aba
3
+ size 1576208368
model.safetensors.index.json ADDED
@@ -0,0 +1,934 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 5575277568
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "model-00002-of-00002.safetensors",
7
+ "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
8
+ "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
9
+ "model.layers.0.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
10
+ "model.layers.0.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
11
+ "model.layers.0.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
12
+ "model.layers.0.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
13
+ "model.layers.0.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
14
+ "model.layers.0.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
15
+ "model.layers.0.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
16
+ "model.layers.0.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
17
+ "model.layers.0.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
18
+ "model.layers.0.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
19
+ "model.layers.0.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
20
+ "model.layers.0.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
21
+ "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
22
+ "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
23
+ "model.layers.0.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
24
+ "model.layers.0.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
25
+ "model.layers.0.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
26
+ "model.layers.0.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
27
+ "model.layers.0.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
28
+ "model.layers.0.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
29
+ "model.layers.0.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
30
+ "model.layers.0.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
31
+ "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
32
+ "model.layers.0.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
33
+ "model.layers.0.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
34
+ "model.layers.0.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
35
+ "model.layers.0.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
36
+ "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
37
+ "model.layers.0.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
38
+ "model.layers.0.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
39
+ "model.layers.0.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
40
+ "model.layers.0.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
41
+ "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
42
+ "model.layers.1.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
43
+ "model.layers.1.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
44
+ "model.layers.1.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
45
+ "model.layers.1.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
46
+ "model.layers.1.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
47
+ "model.layers.1.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
48
+ "model.layers.1.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
49
+ "model.layers.1.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
50
+ "model.layers.1.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
51
+ "model.layers.1.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
52
+ "model.layers.1.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
53
+ "model.layers.1.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
54
+ "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
55
+ "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
56
+ "model.layers.1.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
57
+ "model.layers.1.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
58
+ "model.layers.1.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
59
+ "model.layers.1.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
60
+ "model.layers.1.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
61
+ "model.layers.1.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
62
+ "model.layers.1.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
63
+ "model.layers.1.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
64
+ "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
65
+ "model.layers.1.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
66
+ "model.layers.1.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
67
+ "model.layers.1.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
68
+ "model.layers.1.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
69
+ "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
70
+ "model.layers.1.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
71
+ "model.layers.1.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
72
+ "model.layers.1.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
73
+ "model.layers.1.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
74
+ "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
75
+ "model.layers.10.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
76
+ "model.layers.10.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
77
+ "model.layers.10.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
78
+ "model.layers.10.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
79
+ "model.layers.10.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
80
+ "model.layers.10.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
81
+ "model.layers.10.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
82
+ "model.layers.10.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
83
+ "model.layers.10.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
84
+ "model.layers.10.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
85
+ "model.layers.10.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
86
+ "model.layers.10.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
87
+ "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
88
+ "model.layers.10.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
89
+ "model.layers.10.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
90
+ "model.layers.10.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
91
+ "model.layers.10.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
92
+ "model.layers.10.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
93
+ "model.layers.10.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
94
+ "model.layers.10.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
95
+ "model.layers.10.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
96
+ "model.layers.10.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
97
+ "model.layers.10.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
98
+ "model.layers.10.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
99
+ "model.layers.10.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
100
+ "model.layers.10.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
101
+ "model.layers.10.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
102
+ "model.layers.10.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
103
+ "model.layers.10.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
104
+ "model.layers.10.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
105
+ "model.layers.10.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
106
+ "model.layers.10.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
107
+ "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
108
+ "model.layers.11.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
109
+ "model.layers.11.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
110
+ "model.layers.11.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
111
+ "model.layers.11.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
112
+ "model.layers.11.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
113
+ "model.layers.11.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
114
+ "model.layers.11.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
115
+ "model.layers.11.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
116
+ "model.layers.11.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
117
+ "model.layers.11.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
118
+ "model.layers.11.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
119
+ "model.layers.11.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
120
+ "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
121
+ "model.layers.11.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
122
+ "model.layers.11.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
123
+ "model.layers.11.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
124
+ "model.layers.11.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
125
+ "model.layers.11.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
126
+ "model.layers.11.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
127
+ "model.layers.11.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
128
+ "model.layers.11.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
129
+ "model.layers.11.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
130
+ "model.layers.11.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
131
+ "model.layers.11.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
132
+ "model.layers.11.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
133
+ "model.layers.11.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
134
+ "model.layers.11.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
135
+ "model.layers.11.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
136
+ "model.layers.11.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
137
+ "model.layers.11.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
138
+ "model.layers.11.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
139
+ "model.layers.11.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
140
+ "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
141
+ "model.layers.12.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
142
+ "model.layers.12.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
143
+ "model.layers.12.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
144
+ "model.layers.12.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
145
+ "model.layers.12.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
146
+ "model.layers.12.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
147
+ "model.layers.12.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
148
+ "model.layers.12.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
149
+ "model.layers.12.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
150
+ "model.layers.12.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
151
+ "model.layers.12.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
152
+ "model.layers.12.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
153
+ "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
154
+ "model.layers.12.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
155
+ "model.layers.12.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
156
+ "model.layers.12.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
157
+ "model.layers.12.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
158
+ "model.layers.12.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
159
+ "model.layers.12.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
160
+ "model.layers.12.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
161
+ "model.layers.12.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
162
+ "model.layers.12.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
163
+ "model.layers.12.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
164
+ "model.layers.12.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
165
+ "model.layers.12.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
166
+ "model.layers.12.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
167
+ "model.layers.12.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
168
+ "model.layers.12.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
169
+ "model.layers.12.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
170
+ "model.layers.12.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
171
+ "model.layers.12.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
172
+ "model.layers.12.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
173
+ "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
174
+ "model.layers.13.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
175
+ "model.layers.13.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
176
+ "model.layers.13.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
177
+ "model.layers.13.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
178
+ "model.layers.13.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
179
+ "model.layers.13.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
180
+ "model.layers.13.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
181
+ "model.layers.13.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
182
+ "model.layers.13.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
183
+ "model.layers.13.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
184
+ "model.layers.13.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
185
+ "model.layers.13.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
186
+ "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
187
+ "model.layers.13.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
188
+ "model.layers.13.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
189
+ "model.layers.13.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
190
+ "model.layers.13.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
191
+ "model.layers.13.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
192
+ "model.layers.13.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
193
+ "model.layers.13.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
194
+ "model.layers.13.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
195
+ "model.layers.13.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
196
+ "model.layers.13.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
197
+ "model.layers.13.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
198
+ "model.layers.13.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
199
+ "model.layers.13.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
200
+ "model.layers.13.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
201
+ "model.layers.13.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
202
+ "model.layers.13.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
203
+ "model.layers.13.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
204
+ "model.layers.13.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
205
+ "model.layers.13.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
206
+ "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
207
+ "model.layers.14.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
208
+ "model.layers.14.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
209
+ "model.layers.14.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
210
+ "model.layers.14.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
211
+ "model.layers.14.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
212
+ "model.layers.14.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
213
+ "model.layers.14.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
214
+ "model.layers.14.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
215
+ "model.layers.14.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
216
+ "model.layers.14.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
217
+ "model.layers.14.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
218
+ "model.layers.14.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
219
+ "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
220
+ "model.layers.14.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
221
+ "model.layers.14.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
222
+ "model.layers.14.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
223
+ "model.layers.14.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
224
+ "model.layers.14.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
225
+ "model.layers.14.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
226
+ "model.layers.14.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
227
+ "model.layers.14.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
228
+ "model.layers.14.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
229
+ "model.layers.14.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
230
+ "model.layers.14.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
231
+ "model.layers.14.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
232
+ "model.layers.14.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
233
+ "model.layers.14.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
234
+ "model.layers.14.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
235
+ "model.layers.14.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
236
+ "model.layers.14.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
237
+ "model.layers.14.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
238
+ "model.layers.14.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
239
+ "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
240
+ "model.layers.15.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
241
+ "model.layers.15.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
242
+ "model.layers.15.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
243
+ "model.layers.15.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
244
+ "model.layers.15.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
245
+ "model.layers.15.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
246
+ "model.layers.15.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
247
+ "model.layers.15.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
248
+ "model.layers.15.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
249
+ "model.layers.15.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
250
+ "model.layers.15.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
251
+ "model.layers.15.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
252
+ "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
253
+ "model.layers.15.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
254
+ "model.layers.15.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
255
+ "model.layers.15.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
256
+ "model.layers.15.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
257
+ "model.layers.15.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
258
+ "model.layers.15.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
259
+ "model.layers.15.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
260
+ "model.layers.15.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
261
+ "model.layers.15.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
262
+ "model.layers.15.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
263
+ "model.layers.15.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
264
+ "model.layers.15.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
265
+ "model.layers.15.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
266
+ "model.layers.15.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
267
+ "model.layers.15.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
268
+ "model.layers.15.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
269
+ "model.layers.15.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
270
+ "model.layers.15.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
271
+ "model.layers.15.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
272
+ "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
273
+ "model.layers.16.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
274
+ "model.layers.16.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
275
+ "model.layers.16.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
276
+ "model.layers.16.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
277
+ "model.layers.16.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
278
+ "model.layers.16.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
279
+ "model.layers.16.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
280
+ "model.layers.16.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
281
+ "model.layers.16.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
282
+ "model.layers.16.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
283
+ "model.layers.16.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
284
+ "model.layers.16.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
285
+ "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
286
+ "model.layers.16.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
287
+ "model.layers.16.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
288
+ "model.layers.16.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
289
+ "model.layers.16.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
290
+ "model.layers.16.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
291
+ "model.layers.16.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
292
+ "model.layers.16.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
293
+ "model.layers.16.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
294
+ "model.layers.16.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
295
+ "model.layers.16.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
296
+ "model.layers.16.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
297
+ "model.layers.16.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
298
+ "model.layers.16.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
299
+ "model.layers.16.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
300
+ "model.layers.16.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
301
+ "model.layers.16.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
302
+ "model.layers.16.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
303
+ "model.layers.16.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
304
+ "model.layers.16.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
305
+ "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
306
+ "model.layers.17.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
307
+ "model.layers.17.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
308
+ "model.layers.17.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
309
+ "model.layers.17.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
310
+ "model.layers.17.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
311
+ "model.layers.17.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
312
+ "model.layers.17.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
313
+ "model.layers.17.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
314
+ "model.layers.17.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
315
+ "model.layers.17.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
316
+ "model.layers.17.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
317
+ "model.layers.17.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
318
+ "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
319
+ "model.layers.17.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
320
+ "model.layers.17.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
321
+ "model.layers.17.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
322
+ "model.layers.17.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
323
+ "model.layers.17.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
324
+ "model.layers.17.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
325
+ "model.layers.17.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
326
+ "model.layers.17.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
327
+ "model.layers.17.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
328
+ "model.layers.17.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
329
+ "model.layers.17.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
330
+ "model.layers.17.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
331
+ "model.layers.17.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
332
+ "model.layers.17.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
333
+ "model.layers.17.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
334
+ "model.layers.17.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
335
+ "model.layers.17.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
336
+ "model.layers.17.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
337
+ "model.layers.17.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
338
+ "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
339
+ "model.layers.18.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
340
+ "model.layers.18.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
341
+ "model.layers.18.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
342
+ "model.layers.18.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
343
+ "model.layers.18.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
344
+ "model.layers.18.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
345
+ "model.layers.18.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
346
+ "model.layers.18.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
347
+ "model.layers.18.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
348
+ "model.layers.18.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
349
+ "model.layers.18.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
350
+ "model.layers.18.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
351
+ "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
352
+ "model.layers.18.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
353
+ "model.layers.18.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
354
+ "model.layers.18.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
355
+ "model.layers.18.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
356
+ "model.layers.18.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
357
+ "model.layers.18.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
358
+ "model.layers.18.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
359
+ "model.layers.18.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
360
+ "model.layers.18.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
361
+ "model.layers.18.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
362
+ "model.layers.18.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
363
+ "model.layers.18.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
364
+ "model.layers.18.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
365
+ "model.layers.18.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
366
+ "model.layers.18.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
367
+ "model.layers.18.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
368
+ "model.layers.18.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
369
+ "model.layers.18.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
370
+ "model.layers.18.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
371
+ "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
372
+ "model.layers.19.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
373
+ "model.layers.19.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
374
+ "model.layers.19.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
375
+ "model.layers.19.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
376
+ "model.layers.19.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
377
+ "model.layers.19.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
378
+ "model.layers.19.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
379
+ "model.layers.19.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
380
+ "model.layers.19.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
381
+ "model.layers.19.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
382
+ "model.layers.19.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
383
+ "model.layers.19.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
384
+ "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
385
+ "model.layers.19.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
386
+ "model.layers.19.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
387
+ "model.layers.19.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
388
+ "model.layers.19.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
389
+ "model.layers.19.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
390
+ "model.layers.19.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
391
+ "model.layers.19.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
392
+ "model.layers.19.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
393
+ "model.layers.19.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
394
+ "model.layers.19.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
395
+ "model.layers.19.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
396
+ "model.layers.19.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
397
+ "model.layers.19.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
398
+ "model.layers.19.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
399
+ "model.layers.19.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
400
+ "model.layers.19.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
401
+ "model.layers.19.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
402
+ "model.layers.19.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
403
+ "model.layers.19.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
404
+ "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
405
+ "model.layers.2.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
406
+ "model.layers.2.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
407
+ "model.layers.2.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
408
+ "model.layers.2.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
409
+ "model.layers.2.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
410
+ "model.layers.2.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
411
+ "model.layers.2.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
412
+ "model.layers.2.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
413
+ "model.layers.2.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
414
+ "model.layers.2.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
415
+ "model.layers.2.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
416
+ "model.layers.2.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
417
+ "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
418
+ "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
419
+ "model.layers.2.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
420
+ "model.layers.2.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
421
+ "model.layers.2.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
422
+ "model.layers.2.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
423
+ "model.layers.2.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
424
+ "model.layers.2.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
425
+ "model.layers.2.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
426
+ "model.layers.2.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
427
+ "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
428
+ "model.layers.2.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
429
+ "model.layers.2.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
430
+ "model.layers.2.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
431
+ "model.layers.2.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
432
+ "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
433
+ "model.layers.2.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
434
+ "model.layers.2.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
435
+ "model.layers.2.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
436
+ "model.layers.2.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
437
+ "model.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
438
+ "model.layers.20.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
439
+ "model.layers.20.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
440
+ "model.layers.20.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
441
+ "model.layers.20.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
442
+ "model.layers.20.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
443
+ "model.layers.20.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
444
+ "model.layers.20.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
445
+ "model.layers.20.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
446
+ "model.layers.20.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
447
+ "model.layers.20.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
448
+ "model.layers.20.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
449
+ "model.layers.20.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
450
+ "model.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
451
+ "model.layers.20.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
452
+ "model.layers.20.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
453
+ "model.layers.20.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
454
+ "model.layers.20.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
455
+ "model.layers.20.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
456
+ "model.layers.20.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
457
+ "model.layers.20.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
458
+ "model.layers.20.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
459
+ "model.layers.20.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
460
+ "model.layers.20.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
461
+ "model.layers.20.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
462
+ "model.layers.20.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
463
+ "model.layers.20.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
464
+ "model.layers.20.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
465
+ "model.layers.20.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
466
+ "model.layers.20.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
467
+ "model.layers.20.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
468
+ "model.layers.20.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
469
+ "model.layers.20.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
470
+ "model.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
471
+ "model.layers.21.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
472
+ "model.layers.21.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
473
+ "model.layers.21.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
474
+ "model.layers.21.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
475
+ "model.layers.21.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
476
+ "model.layers.21.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
477
+ "model.layers.21.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
478
+ "model.layers.21.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
479
+ "model.layers.21.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
480
+ "model.layers.21.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
481
+ "model.layers.21.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
482
+ "model.layers.21.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
483
+ "model.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
484
+ "model.layers.21.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
485
+ "model.layers.21.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
486
+ "model.layers.21.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
487
+ "model.layers.21.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
488
+ "model.layers.21.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
489
+ "model.layers.21.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
490
+ "model.layers.21.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
491
+ "model.layers.21.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
492
+ "model.layers.21.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
493
+ "model.layers.21.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
494
+ "model.layers.21.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
495
+ "model.layers.21.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
496
+ "model.layers.21.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
497
+ "model.layers.21.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
498
+ "model.layers.21.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
499
+ "model.layers.21.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
500
+ "model.layers.21.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
501
+ "model.layers.21.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
502
+ "model.layers.21.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
503
+ "model.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
504
+ "model.layers.22.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
505
+ "model.layers.22.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
506
+ "model.layers.22.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
507
+ "model.layers.22.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
508
+ "model.layers.22.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
509
+ "model.layers.22.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
510
+ "model.layers.22.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
511
+ "model.layers.22.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
512
+ "model.layers.22.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
513
+ "model.layers.22.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
514
+ "model.layers.22.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
515
+ "model.layers.22.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
516
+ "model.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
517
+ "model.layers.22.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
518
+ "model.layers.22.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
519
+ "model.layers.22.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
520
+ "model.layers.22.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
521
+ "model.layers.22.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
522
+ "model.layers.22.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
523
+ "model.layers.22.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
524
+ "model.layers.22.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
525
+ "model.layers.22.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
526
+ "model.layers.22.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
527
+ "model.layers.22.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
528
+ "model.layers.22.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
529
+ "model.layers.22.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
530
+ "model.layers.22.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
531
+ "model.layers.22.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
532
+ "model.layers.22.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
533
+ "model.layers.22.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
534
+ "model.layers.22.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
535
+ "model.layers.22.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
536
+ "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
537
+ "model.layers.23.mlp.down_proj.g_idx": "model-00002-of-00002.safetensors",
538
+ "model.layers.23.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
539
+ "model.layers.23.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
540
+ "model.layers.23.mlp.down_proj.scales": "model-00002-of-00002.safetensors",
541
+ "model.layers.23.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
542
+ "model.layers.23.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
543
+ "model.layers.23.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
544
+ "model.layers.23.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
545
+ "model.layers.23.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
546
+ "model.layers.23.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
547
+ "model.layers.23.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
548
+ "model.layers.23.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
549
+ "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
550
+ "model.layers.23.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
551
+ "model.layers.23.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
552
+ "model.layers.23.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
553
+ "model.layers.23.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
554
+ "model.layers.23.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
555
+ "model.layers.23.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
556
+ "model.layers.23.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
557
+ "model.layers.23.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
558
+ "model.layers.23.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
559
+ "model.layers.23.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
560
+ "model.layers.23.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
561
+ "model.layers.23.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
562
+ "model.layers.23.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
563
+ "model.layers.23.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
564
+ "model.layers.23.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
565
+ "model.layers.23.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
566
+ "model.layers.23.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
567
+ "model.layers.23.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
568
+ "model.layers.23.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
569
+ "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
570
+ "model.layers.24.mlp.down_proj.g_idx": "model-00002-of-00002.safetensors",
571
+ "model.layers.24.mlp.down_proj.qweight": "model-00002-of-00002.safetensors",
572
+ "model.layers.24.mlp.down_proj.qzeros": "model-00002-of-00002.safetensors",
573
+ "model.layers.24.mlp.down_proj.scales": "model-00002-of-00002.safetensors",
574
+ "model.layers.24.mlp.gate_proj.g_idx": "model-00002-of-00002.safetensors",
575
+ "model.layers.24.mlp.gate_proj.qweight": "model-00002-of-00002.safetensors",
576
+ "model.layers.24.mlp.gate_proj.qzeros": "model-00002-of-00002.safetensors",
577
+ "model.layers.24.mlp.gate_proj.scales": "model-00002-of-00002.safetensors",
578
+ "model.layers.24.mlp.up_proj.g_idx": "model-00002-of-00002.safetensors",
579
+ "model.layers.24.mlp.up_proj.qweight": "model-00002-of-00002.safetensors",
580
+ "model.layers.24.mlp.up_proj.qzeros": "model-00002-of-00002.safetensors",
581
+ "model.layers.24.mlp.up_proj.scales": "model-00002-of-00002.safetensors",
582
+ "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
583
+ "model.layers.24.self_attn.k_proj.bias": "model-00002-of-00002.safetensors",
584
+ "model.layers.24.self_attn.k_proj.g_idx": "model-00002-of-00002.safetensors",
585
+ "model.layers.24.self_attn.k_proj.qweight": "model-00002-of-00002.safetensors",
586
+ "model.layers.24.self_attn.k_proj.qzeros": "model-00002-of-00002.safetensors",
587
+ "model.layers.24.self_attn.k_proj.scales": "model-00002-of-00002.safetensors",
588
+ "model.layers.24.self_attn.o_proj.g_idx": "model-00002-of-00002.safetensors",
589
+ "model.layers.24.self_attn.o_proj.qweight": "model-00002-of-00002.safetensors",
590
+ "model.layers.24.self_attn.o_proj.qzeros": "model-00002-of-00002.safetensors",
591
+ "model.layers.24.self_attn.o_proj.scales": "model-00002-of-00002.safetensors",
592
+ "model.layers.24.self_attn.q_proj.bias": "model-00002-of-00002.safetensors",
593
+ "model.layers.24.self_attn.q_proj.g_idx": "model-00002-of-00002.safetensors",
594
+ "model.layers.24.self_attn.q_proj.qweight": "model-00002-of-00002.safetensors",
595
+ "model.layers.24.self_attn.q_proj.qzeros": "model-00002-of-00002.safetensors",
596
+ "model.layers.24.self_attn.q_proj.scales": "model-00002-of-00002.safetensors",
597
+ "model.layers.24.self_attn.v_proj.bias": "model-00002-of-00002.safetensors",
598
+ "model.layers.24.self_attn.v_proj.g_idx": "model-00002-of-00002.safetensors",
599
+ "model.layers.24.self_attn.v_proj.qweight": "model-00002-of-00002.safetensors",
600
+ "model.layers.24.self_attn.v_proj.qzeros": "model-00002-of-00002.safetensors",
601
+ "model.layers.24.self_attn.v_proj.scales": "model-00002-of-00002.safetensors",
602
+ "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
603
+ "model.layers.25.mlp.down_proj.g_idx": "model-00002-of-00002.safetensors",
604
+ "model.layers.25.mlp.down_proj.qweight": "model-00002-of-00002.safetensors",
605
+ "model.layers.25.mlp.down_proj.qzeros": "model-00002-of-00002.safetensors",
606
+ "model.layers.25.mlp.down_proj.scales": "model-00002-of-00002.safetensors",
607
+ "model.layers.25.mlp.gate_proj.g_idx": "model-00002-of-00002.safetensors",
608
+ "model.layers.25.mlp.gate_proj.qweight": "model-00002-of-00002.safetensors",
609
+ "model.layers.25.mlp.gate_proj.qzeros": "model-00002-of-00002.safetensors",
610
+ "model.layers.25.mlp.gate_proj.scales": "model-00002-of-00002.safetensors",
611
+ "model.layers.25.mlp.up_proj.g_idx": "model-00002-of-00002.safetensors",
612
+ "model.layers.25.mlp.up_proj.qweight": "model-00002-of-00002.safetensors",
613
+ "model.layers.25.mlp.up_proj.qzeros": "model-00002-of-00002.safetensors",
614
+ "model.layers.25.mlp.up_proj.scales": "model-00002-of-00002.safetensors",
615
+ "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
616
+ "model.layers.25.self_attn.k_proj.bias": "model-00002-of-00002.safetensors",
617
+ "model.layers.25.self_attn.k_proj.g_idx": "model-00002-of-00002.safetensors",
618
+ "model.layers.25.self_attn.k_proj.qweight": "model-00002-of-00002.safetensors",
619
+ "model.layers.25.self_attn.k_proj.qzeros": "model-00002-of-00002.safetensors",
620
+ "model.layers.25.self_attn.k_proj.scales": "model-00002-of-00002.safetensors",
621
+ "model.layers.25.self_attn.o_proj.g_idx": "model-00002-of-00002.safetensors",
622
+ "model.layers.25.self_attn.o_proj.qweight": "model-00002-of-00002.safetensors",
623
+ "model.layers.25.self_attn.o_proj.qzeros": "model-00002-of-00002.safetensors",
624
+ "model.layers.25.self_attn.o_proj.scales": "model-00002-of-00002.safetensors",
625
+ "model.layers.25.self_attn.q_proj.bias": "model-00002-of-00002.safetensors",
626
+ "model.layers.25.self_attn.q_proj.g_idx": "model-00002-of-00002.safetensors",
627
+ "model.layers.25.self_attn.q_proj.qweight": "model-00002-of-00002.safetensors",
628
+ "model.layers.25.self_attn.q_proj.qzeros": "model-00002-of-00002.safetensors",
629
+ "model.layers.25.self_attn.q_proj.scales": "model-00002-of-00002.safetensors",
630
+ "model.layers.25.self_attn.v_proj.bias": "model-00002-of-00002.safetensors",
631
+ "model.layers.25.self_attn.v_proj.g_idx": "model-00002-of-00002.safetensors",
632
+ "model.layers.25.self_attn.v_proj.qweight": "model-00002-of-00002.safetensors",
633
+ "model.layers.25.self_attn.v_proj.qzeros": "model-00002-of-00002.safetensors",
634
+ "model.layers.25.self_attn.v_proj.scales": "model-00002-of-00002.safetensors",
635
+ "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
636
+ "model.layers.26.mlp.down_proj.g_idx": "model-00002-of-00002.safetensors",
637
+ "model.layers.26.mlp.down_proj.qweight": "model-00002-of-00002.safetensors",
638
+ "model.layers.26.mlp.down_proj.qzeros": "model-00002-of-00002.safetensors",
639
+ "model.layers.26.mlp.down_proj.scales": "model-00002-of-00002.safetensors",
640
+ "model.layers.26.mlp.gate_proj.g_idx": "model-00002-of-00002.safetensors",
641
+ "model.layers.26.mlp.gate_proj.qweight": "model-00002-of-00002.safetensors",
642
+ "model.layers.26.mlp.gate_proj.qzeros": "model-00002-of-00002.safetensors",
643
+ "model.layers.26.mlp.gate_proj.scales": "model-00002-of-00002.safetensors",
644
+ "model.layers.26.mlp.up_proj.g_idx": "model-00002-of-00002.safetensors",
645
+ "model.layers.26.mlp.up_proj.qweight": "model-00002-of-00002.safetensors",
646
+ "model.layers.26.mlp.up_proj.qzeros": "model-00002-of-00002.safetensors",
647
+ "model.layers.26.mlp.up_proj.scales": "model-00002-of-00002.safetensors",
648
+ "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
649
+ "model.layers.26.self_attn.k_proj.bias": "model-00002-of-00002.safetensors",
650
+ "model.layers.26.self_attn.k_proj.g_idx": "model-00002-of-00002.safetensors",
651
+ "model.layers.26.self_attn.k_proj.qweight": "model-00002-of-00002.safetensors",
652
+ "model.layers.26.self_attn.k_proj.qzeros": "model-00002-of-00002.safetensors",
653
+ "model.layers.26.self_attn.k_proj.scales": "model-00002-of-00002.safetensors",
654
+ "model.layers.26.self_attn.o_proj.g_idx": "model-00002-of-00002.safetensors",
655
+ "model.layers.26.self_attn.o_proj.qweight": "model-00002-of-00002.safetensors",
656
+ "model.layers.26.self_attn.o_proj.qzeros": "model-00002-of-00002.safetensors",
657
+ "model.layers.26.self_attn.o_proj.scales": "model-00002-of-00002.safetensors",
658
+ "model.layers.26.self_attn.q_proj.bias": "model-00002-of-00002.safetensors",
659
+ "model.layers.26.self_attn.q_proj.g_idx": "model-00002-of-00002.safetensors",
660
+ "model.layers.26.self_attn.q_proj.qweight": "model-00002-of-00002.safetensors",
661
+ "model.layers.26.self_attn.q_proj.qzeros": "model-00002-of-00002.safetensors",
662
+ "model.layers.26.self_attn.q_proj.scales": "model-00002-of-00002.safetensors",
663
+ "model.layers.26.self_attn.v_proj.bias": "model-00002-of-00002.safetensors",
664
+ "model.layers.26.self_attn.v_proj.g_idx": "model-00002-of-00002.safetensors",
665
+ "model.layers.26.self_attn.v_proj.qweight": "model-00002-of-00002.safetensors",
666
+ "model.layers.26.self_attn.v_proj.qzeros": "model-00002-of-00002.safetensors",
667
+ "model.layers.26.self_attn.v_proj.scales": "model-00002-of-00002.safetensors",
668
+ "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
669
+ "model.layers.27.mlp.down_proj.g_idx": "model-00002-of-00002.safetensors",
670
+ "model.layers.27.mlp.down_proj.qweight": "model-00002-of-00002.safetensors",
671
+ "model.layers.27.mlp.down_proj.qzeros": "model-00002-of-00002.safetensors",
672
+ "model.layers.27.mlp.down_proj.scales": "model-00002-of-00002.safetensors",
673
+ "model.layers.27.mlp.gate_proj.g_idx": "model-00002-of-00002.safetensors",
674
+ "model.layers.27.mlp.gate_proj.qweight": "model-00002-of-00002.safetensors",
675
+ "model.layers.27.mlp.gate_proj.qzeros": "model-00002-of-00002.safetensors",
676
+ "model.layers.27.mlp.gate_proj.scales": "model-00002-of-00002.safetensors",
677
+ "model.layers.27.mlp.up_proj.g_idx": "model-00002-of-00002.safetensors",
678
+ "model.layers.27.mlp.up_proj.qweight": "model-00002-of-00002.safetensors",
679
+ "model.layers.27.mlp.up_proj.qzeros": "model-00002-of-00002.safetensors",
680
+ "model.layers.27.mlp.up_proj.scales": "model-00002-of-00002.safetensors",
681
+ "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
682
+ "model.layers.27.self_attn.k_proj.bias": "model-00002-of-00002.safetensors",
683
+ "model.layers.27.self_attn.k_proj.g_idx": "model-00002-of-00002.safetensors",
684
+ "model.layers.27.self_attn.k_proj.qweight": "model-00002-of-00002.safetensors",
685
+ "model.layers.27.self_attn.k_proj.qzeros": "model-00002-of-00002.safetensors",
686
+ "model.layers.27.self_attn.k_proj.scales": "model-00002-of-00002.safetensors",
687
+ "model.layers.27.self_attn.o_proj.g_idx": "model-00002-of-00002.safetensors",
688
+ "model.layers.27.self_attn.o_proj.qweight": "model-00002-of-00002.safetensors",
689
+ "model.layers.27.self_attn.o_proj.qzeros": "model-00002-of-00002.safetensors",
690
+ "model.layers.27.self_attn.o_proj.scales": "model-00002-of-00002.safetensors",
691
+ "model.layers.27.self_attn.q_proj.bias": "model-00002-of-00002.safetensors",
692
+ "model.layers.27.self_attn.q_proj.g_idx": "model-00002-of-00002.safetensors",
693
+ "model.layers.27.self_attn.q_proj.qweight": "model-00002-of-00002.safetensors",
694
+ "model.layers.27.self_attn.q_proj.qzeros": "model-00002-of-00002.safetensors",
695
+ "model.layers.27.self_attn.q_proj.scales": "model-00002-of-00002.safetensors",
696
+ "model.layers.27.self_attn.v_proj.bias": "model-00002-of-00002.safetensors",
697
+ "model.layers.27.self_attn.v_proj.g_idx": "model-00002-of-00002.safetensors",
698
+ "model.layers.27.self_attn.v_proj.qweight": "model-00002-of-00002.safetensors",
699
+ "model.layers.27.self_attn.v_proj.qzeros": "model-00002-of-00002.safetensors",
700
+ "model.layers.27.self_attn.v_proj.scales": "model-00002-of-00002.safetensors",
701
+ "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
702
+ "model.layers.3.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
703
+ "model.layers.3.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
704
+ "model.layers.3.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
705
+ "model.layers.3.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
706
+ "model.layers.3.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
707
+ "model.layers.3.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
708
+ "model.layers.3.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
709
+ "model.layers.3.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
710
+ "model.layers.3.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
711
+ "model.layers.3.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
712
+ "model.layers.3.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
713
+ "model.layers.3.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
714
+ "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
715
+ "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
716
+ "model.layers.3.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
717
+ "model.layers.3.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
718
+ "model.layers.3.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
719
+ "model.layers.3.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
720
+ "model.layers.3.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
721
+ "model.layers.3.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
722
+ "model.layers.3.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
723
+ "model.layers.3.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
724
+ "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
725
+ "model.layers.3.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
726
+ "model.layers.3.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
727
+ "model.layers.3.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
728
+ "model.layers.3.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
729
+ "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
730
+ "model.layers.3.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
731
+ "model.layers.3.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
732
+ "model.layers.3.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
733
+ "model.layers.3.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
734
+ "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
735
+ "model.layers.4.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
736
+ "model.layers.4.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
737
+ "model.layers.4.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
738
+ "model.layers.4.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
739
+ "model.layers.4.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
740
+ "model.layers.4.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
741
+ "model.layers.4.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
742
+ "model.layers.4.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
743
+ "model.layers.4.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
744
+ "model.layers.4.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
745
+ "model.layers.4.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
746
+ "model.layers.4.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
747
+ "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
748
+ "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
749
+ "model.layers.4.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
750
+ "model.layers.4.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
751
+ "model.layers.4.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
752
+ "model.layers.4.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
753
+ "model.layers.4.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
754
+ "model.layers.4.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
755
+ "model.layers.4.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
756
+ "model.layers.4.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
757
+ "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
758
+ "model.layers.4.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
759
+ "model.layers.4.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
760
+ "model.layers.4.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
761
+ "model.layers.4.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
762
+ "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
763
+ "model.layers.4.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
764
+ "model.layers.4.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
765
+ "model.layers.4.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
766
+ "model.layers.4.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
767
+ "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
768
+ "model.layers.5.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
769
+ "model.layers.5.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
770
+ "model.layers.5.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
771
+ "model.layers.5.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
772
+ "model.layers.5.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
773
+ "model.layers.5.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
774
+ "model.layers.5.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
775
+ "model.layers.5.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
776
+ "model.layers.5.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
777
+ "model.layers.5.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
778
+ "model.layers.5.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
779
+ "model.layers.5.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
780
+ "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
781
+ "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
782
+ "model.layers.5.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
783
+ "model.layers.5.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
784
+ "model.layers.5.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
785
+ "model.layers.5.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
786
+ "model.layers.5.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
787
+ "model.layers.5.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
788
+ "model.layers.5.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
789
+ "model.layers.5.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
790
+ "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
791
+ "model.layers.5.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
792
+ "model.layers.5.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
793
+ "model.layers.5.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
794
+ "model.layers.5.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
795
+ "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
796
+ "model.layers.5.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
797
+ "model.layers.5.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
798
+ "model.layers.5.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
799
+ "model.layers.5.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
800
+ "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
801
+ "model.layers.6.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
802
+ "model.layers.6.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
803
+ "model.layers.6.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
804
+ "model.layers.6.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
805
+ "model.layers.6.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
806
+ "model.layers.6.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
807
+ "model.layers.6.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
808
+ "model.layers.6.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
809
+ "model.layers.6.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
810
+ "model.layers.6.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
811
+ "model.layers.6.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
812
+ "model.layers.6.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
813
+ "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
814
+ "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
815
+ "model.layers.6.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
816
+ "model.layers.6.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
817
+ "model.layers.6.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
818
+ "model.layers.6.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
819
+ "model.layers.6.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
820
+ "model.layers.6.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
821
+ "model.layers.6.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
822
+ "model.layers.6.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
823
+ "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
824
+ "model.layers.6.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
825
+ "model.layers.6.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
826
+ "model.layers.6.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
827
+ "model.layers.6.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
828
+ "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
829
+ "model.layers.6.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
830
+ "model.layers.6.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
831
+ "model.layers.6.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
832
+ "model.layers.6.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
833
+ "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
834
+ "model.layers.7.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
835
+ "model.layers.7.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
836
+ "model.layers.7.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
837
+ "model.layers.7.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
838
+ "model.layers.7.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
839
+ "model.layers.7.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
840
+ "model.layers.7.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
841
+ "model.layers.7.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
842
+ "model.layers.7.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
843
+ "model.layers.7.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
844
+ "model.layers.7.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
845
+ "model.layers.7.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
846
+ "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
847
+ "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
848
+ "model.layers.7.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
849
+ "model.layers.7.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
850
+ "model.layers.7.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
851
+ "model.layers.7.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
852
+ "model.layers.7.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
853
+ "model.layers.7.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
854
+ "model.layers.7.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
855
+ "model.layers.7.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
856
+ "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
857
+ "model.layers.7.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
858
+ "model.layers.7.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
859
+ "model.layers.7.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
860
+ "model.layers.7.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
861
+ "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
862
+ "model.layers.7.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
863
+ "model.layers.7.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
864
+ "model.layers.7.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
865
+ "model.layers.7.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
866
+ "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
867
+ "model.layers.8.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
868
+ "model.layers.8.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
869
+ "model.layers.8.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
870
+ "model.layers.8.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
871
+ "model.layers.8.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
872
+ "model.layers.8.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
873
+ "model.layers.8.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
874
+ "model.layers.8.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
875
+ "model.layers.8.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
876
+ "model.layers.8.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
877
+ "model.layers.8.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
878
+ "model.layers.8.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
879
+ "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
880
+ "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
881
+ "model.layers.8.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
882
+ "model.layers.8.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
883
+ "model.layers.8.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
884
+ "model.layers.8.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
885
+ "model.layers.8.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
886
+ "model.layers.8.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
887
+ "model.layers.8.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
888
+ "model.layers.8.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
889
+ "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
890
+ "model.layers.8.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
891
+ "model.layers.8.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
892
+ "model.layers.8.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
893
+ "model.layers.8.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
894
+ "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
895
+ "model.layers.8.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
896
+ "model.layers.8.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
897
+ "model.layers.8.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
898
+ "model.layers.8.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
899
+ "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
900
+ "model.layers.9.mlp.down_proj.g_idx": "model-00001-of-00002.safetensors",
901
+ "model.layers.9.mlp.down_proj.qweight": "model-00001-of-00002.safetensors",
902
+ "model.layers.9.mlp.down_proj.qzeros": "model-00001-of-00002.safetensors",
903
+ "model.layers.9.mlp.down_proj.scales": "model-00001-of-00002.safetensors",
904
+ "model.layers.9.mlp.gate_proj.g_idx": "model-00001-of-00002.safetensors",
905
+ "model.layers.9.mlp.gate_proj.qweight": "model-00001-of-00002.safetensors",
906
+ "model.layers.9.mlp.gate_proj.qzeros": "model-00001-of-00002.safetensors",
907
+ "model.layers.9.mlp.gate_proj.scales": "model-00001-of-00002.safetensors",
908
+ "model.layers.9.mlp.up_proj.g_idx": "model-00001-of-00002.safetensors",
909
+ "model.layers.9.mlp.up_proj.qweight": "model-00001-of-00002.safetensors",
910
+ "model.layers.9.mlp.up_proj.qzeros": "model-00001-of-00002.safetensors",
911
+ "model.layers.9.mlp.up_proj.scales": "model-00001-of-00002.safetensors",
912
+ "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
913
+ "model.layers.9.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
914
+ "model.layers.9.self_attn.k_proj.g_idx": "model-00001-of-00002.safetensors",
915
+ "model.layers.9.self_attn.k_proj.qweight": "model-00001-of-00002.safetensors",
916
+ "model.layers.9.self_attn.k_proj.qzeros": "model-00001-of-00002.safetensors",
917
+ "model.layers.9.self_attn.k_proj.scales": "model-00001-of-00002.safetensors",
918
+ "model.layers.9.self_attn.o_proj.g_idx": "model-00001-of-00002.safetensors",
919
+ "model.layers.9.self_attn.o_proj.qweight": "model-00001-of-00002.safetensors",
920
+ "model.layers.9.self_attn.o_proj.qzeros": "model-00001-of-00002.safetensors",
921
+ "model.layers.9.self_attn.o_proj.scales": "model-00001-of-00002.safetensors",
922
+ "model.layers.9.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
923
+ "model.layers.9.self_attn.q_proj.g_idx": "model-00001-of-00002.safetensors",
924
+ "model.layers.9.self_attn.q_proj.qweight": "model-00001-of-00002.safetensors",
925
+ "model.layers.9.self_attn.q_proj.qzeros": "model-00001-of-00002.safetensors",
926
+ "model.layers.9.self_attn.q_proj.scales": "model-00001-of-00002.safetensors",
927
+ "model.layers.9.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
928
+ "model.layers.9.self_attn.v_proj.g_idx": "model-00001-of-00002.safetensors",
929
+ "model.layers.9.self_attn.v_proj.qweight": "model-00001-of-00002.safetensors",
930
+ "model.layers.9.self_attn.v_proj.qzeros": "model-00001-of-00002.safetensors",
931
+ "model.layers.9.self_attn.v_proj.scales": "model-00001-of-00002.safetensors",
932
+ "model.norm.weight": "model-00002-of-00002.safetensors"
933
+ }
934
+ }
quant_log.csv ADDED
@@ -0,0 +1,197 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ layer,module,loss,samples,damp,time
2
+ 0,self_attn.k_proj,0.00006717,0.01000,0.743
3
+ 0,self_attn.v_proj,0.00000782,0.01000,0.540
4
+ 0,self_attn.q_proj,0.00026253,0.01000,0.537
5
+ 0,self_attn.o_proj,0.00008649,0.01000,0.538
6
+ 0,mlp.up_proj,0.00024544,0.01000,0.564
7
+ 0,mlp.gate_proj,0.00042476,0.01000,0.558
8
+ 0,mlp.down_proj,304.70507812,0.01250,4.287
9
+ 1,self_attn.k_proj,0.00003458,0.01000,0.543
10
+ 1,self_attn.v_proj,0.00001160,0.01000,0.540
11
+ 1,self_attn.q_proj,0.00012922,0.01000,0.539
12
+ 1,self_attn.o_proj,0.00001389,0.01000,0.540
13
+ 1,mlp.up_proj,0.00507102,0.01000,0.564
14
+ 1,mlp.gate_proj,0.00740672,0.01000,0.561
15
+ 1,mlp.down_proj,201.20426941,0.01250,4.277
16
+ 2,self_attn.k_proj,0.00002051,0.01000,0.545
17
+ 2,self_attn.v_proj,0.00000462,0.01000,0.542
18
+ 2,self_attn.q_proj,0.00008908,0.01000,0.541
19
+ 2,self_attn.o_proj,0.00002082,0.01000,0.540
20
+ 2,mlp.up_proj,0.00323103,0.01000,0.564
21
+ 2,mlp.gate_proj,0.00482144,0.01000,0.559
22
+ 2,mlp.down_proj,265.71157837,0.01250,4.258
23
+ 3,self_attn.k_proj,0.00003836,0.01000,0.542
24
+ 3,self_attn.v_proj,0.00001136,0.01000,0.542
25
+ 3,self_attn.q_proj,0.00017052,0.01000,0.537
26
+ 3,self_attn.o_proj,0.00010798,0.01000,0.541
27
+ 3,mlp.up_proj,0.01222713,0.01000,0.562
28
+ 3,mlp.gate_proj,0.01548911,0.01000,0.558
29
+ 3,mlp.down_proj,0.12992305,0.01250,4.263
30
+ 4,self_attn.k_proj,0.00000324,0.01000,0.543
31
+ 4,self_attn.v_proj,0.00000160,0.01000,0.538
32
+ 4,self_attn.q_proj,0.00001570,0.01000,0.539
33
+ 4,self_attn.o_proj,0.00000409,0.01000,0.540
34
+ 4,mlp.up_proj,0.00579107,0.01000,0.564
35
+ 4,mlp.gate_proj,0.00821990,0.01000,0.562
36
+ 4,mlp.down_proj,0.17879479,0.01250,4.295
37
+ 5,self_attn.k_proj,0.00000587,0.01000,0.542
38
+ 5,self_attn.v_proj,0.00000278,0.01000,0.539
39
+ 5,self_attn.q_proj,0.00002992,0.01000,0.543
40
+ 5,self_attn.o_proj,0.00000610,0.01000,0.544
41
+ 5,mlp.up_proj,0.01783431,0.01000,0.566
42
+ 5,mlp.gate_proj,0.02152922,0.01000,0.562
43
+ 5,mlp.down_proj,7.94002533,0.01250,4.278
44
+ 6,self_attn.k_proj,0.00000566,0.01000,0.546
45
+ 6,self_attn.v_proj,0.00000298,0.01000,0.543
46
+ 6,self_attn.q_proj,0.00002906,0.01000,0.545
47
+ 6,self_attn.o_proj,0.00000604,0.01000,0.543
48
+ 6,mlp.up_proj,0.00136981,0.01000,0.563
49
+ 6,mlp.gate_proj,0.00192878,0.01000,0.563
50
+ 6,mlp.down_proj,29.80362320,0.01250,4.302
51
+ 7,self_attn.k_proj,0.00000723,0.01000,0.543
52
+ 7,self_attn.v_proj,0.00000818,0.01000,0.543
53
+ 7,self_attn.q_proj,0.00004528,0.01000,0.540
54
+ 7,self_attn.o_proj,0.00000810,0.01000,0.542
55
+ 7,mlp.up_proj,0.00104278,0.01000,0.565
56
+ 7,mlp.gate_proj,0.00117535,0.01000,0.562
57
+ 7,mlp.down_proj,73.25530243,0.01250,4.280
58
+ 8,self_attn.k_proj,0.00000813,0.01000,0.540
59
+ 8,self_attn.v_proj,0.00000380,0.01000,0.538
60
+ 8,self_attn.q_proj,0.00003767,0.01000,0.538
61
+ 8,self_attn.o_proj,0.00001477,0.01000,0.540
62
+ 8,mlp.up_proj,0.00108770,0.01000,0.566
63
+ 8,mlp.gate_proj,0.00114401,0.01000,0.563
64
+ 8,mlp.down_proj,118.72260284,0.01250,4.273
65
+ 9,self_attn.k_proj,0.00000741,0.01000,0.545
66
+ 9,self_attn.v_proj,0.00000789,0.01000,0.543
67
+ 9,self_attn.q_proj,0.00004401,0.01000,0.542
68
+ 9,self_attn.o_proj,0.00001458,0.01000,0.540
69
+ 9,mlp.up_proj,0.00241637,0.01000,0.562
70
+ 9,mlp.gate_proj,0.00358828,0.01000,0.560
71
+ 9,mlp.down_proj,20.12282372,0.01250,4.280
72
+ 10,self_attn.k_proj,0.00000835,0.01000,0.544
73
+ 10,self_attn.v_proj,0.00000538,0.01000,0.541
74
+ 10,self_attn.q_proj,0.00004422,0.01000,0.542
75
+ 10,self_attn.o_proj,0.00001159,0.01000,0.542
76
+ 10,mlp.up_proj,0.00074800,0.01000,0.565
77
+ 10,mlp.gate_proj,0.00083565,0.01000,0.561
78
+ 10,mlp.down_proj,351.15560913,0.01250,4.264
79
+ 11,self_attn.k_proj,0.00001277,0.01000,0.546
80
+ 11,self_attn.v_proj,0.00000565,0.01000,0.541
81
+ 11,self_attn.q_proj,0.00006044,0.01000,0.540
82
+ 11,self_attn.o_proj,0.00001777,0.01000,0.540
83
+ 11,mlp.up_proj,0.00086147,0.01000,0.564
84
+ 11,mlp.gate_proj,0.00089606,0.01000,0.565
85
+ 11,mlp.down_proj,4.85732031,0.01250,4.274
86
+ 12,self_attn.k_proj,0.00001361,0.01000,0.541
87
+ 12,self_attn.v_proj,0.00000723,0.01000,0.538
88
+ 12,self_attn.q_proj,0.00006449,0.01000,0.538
89
+ 12,self_attn.o_proj,0.00001664,0.01000,0.539
90
+ 12,mlp.up_proj,0.00072561,0.01000,0.563
91
+ 12,mlp.gate_proj,0.00070967,0.01000,0.566
92
+ 12,mlp.down_proj,84.20599365,0.01250,4.295
93
+ 13,self_attn.k_proj,0.00002183,0.01000,0.545
94
+ 13,self_attn.v_proj,0.00001649,0.01000,0.539
95
+ 13,self_attn.q_proj,0.00011990,0.01000,0.540
96
+ 13,self_attn.o_proj,0.00001876,0.01000,0.541
97
+ 13,mlp.up_proj,0.00052857,0.01000,0.566
98
+ 13,mlp.gate_proj,0.00055504,0.01000,0.563
99
+ 13,mlp.down_proj,434.47857666,0.01250,4.278
100
+ 14,self_attn.k_proj,0.00001691,0.01000,0.546
101
+ 14,self_attn.v_proj,0.00000920,0.01000,0.541
102
+ 14,self_attn.q_proj,0.00008929,0.01000,0.539
103
+ 14,self_attn.o_proj,0.00002732,0.01000,0.539
104
+ 14,mlp.up_proj,0.00055212,0.01000,0.566
105
+ 14,mlp.gate_proj,0.00054635,0.01000,0.562
106
+ 14,mlp.down_proj,410.56881714,0.01250,4.284
107
+ 15,self_attn.k_proj,0.00004251,0.01000,0.548
108
+ 15,self_attn.v_proj,0.00002116,0.01000,0.545
109
+ 15,self_attn.q_proj,0.00020286,0.01000,0.546
110
+ 15,self_attn.o_proj,0.00002885,0.01000,0.542
111
+ 15,mlp.up_proj,0.00051129,0.01000,0.566
112
+ 15,mlp.gate_proj,0.00048951,0.01000,0.564
113
+ 15,mlp.down_proj,815.28625488,0.01250,4.290
114
+ 16,self_attn.k_proj,0.00002103,0.01000,0.546
115
+ 16,self_attn.v_proj,0.00001660,0.01000,0.541
116
+ 16,self_attn.q_proj,0.00011751,0.01000,0.541
117
+ 16,self_attn.o_proj,0.00002044,0.01000,0.540
118
+ 16,mlp.up_proj,0.00043137,0.01000,0.574
119
+ 16,mlp.gate_proj,0.00041475,0.01000,0.562
120
+ 16,mlp.down_proj,1187.52331543,0.01250,4.284
121
+ 17,self_attn.k_proj,0.00002322,0.01000,0.547
122
+ 17,self_attn.v_proj,0.00001877,0.01000,0.542
123
+ 17,self_attn.q_proj,0.00013999,0.01000,0.543
124
+ 17,self_attn.o_proj,0.00003512,0.01000,0.543
125
+ 17,mlp.up_proj,0.00046121,0.01000,0.567
126
+ 17,mlp.gate_proj,0.00043173,0.01000,0.565
127
+ 17,mlp.down_proj,972.71276855,0.01250,4.288
128
+ 18,self_attn.k_proj,0.00001954,0.01000,0.549
129
+ 18,self_attn.v_proj,0.00002303,0.01000,0.546
130
+ 18,self_attn.q_proj,0.00011826,0.01000,0.543
131
+ 18,self_attn.o_proj,0.00003064,0.01000,0.544
132
+ 18,mlp.up_proj,0.00046868,0.01000,0.569
133
+ 18,mlp.gate_proj,0.00043091,0.01000,0.563
134
+ 18,mlp.down_proj,463.82571411,0.01250,4.275
135
+ 19,self_attn.k_proj,0.00001841,0.01000,0.543
136
+ 19,self_attn.v_proj,0.00002542,0.01000,0.540
137
+ 19,self_attn.q_proj,0.00012729,0.01000,0.538
138
+ 19,self_attn.o_proj,0.00003190,0.01000,0.538
139
+ 19,mlp.up_proj,0.00048951,0.01000,0.566
140
+ 19,mlp.gate_proj,0.00047259,0.01000,0.562
141
+ 19,mlp.down_proj,430.34570312,0.01250,4.288
142
+ 20,self_attn.k_proj,0.00001788,0.01000,0.546
143
+ 20,self_attn.v_proj,0.00002601,0.01000,0.543
144
+ 20,self_attn.q_proj,0.00011715,0.01000,0.538
145
+ 20,self_attn.o_proj,0.00003250,0.01000,0.539
146
+ 20,mlp.up_proj,0.00057619,0.01000,0.565
147
+ 20,mlp.gate_proj,0.00055857,0.01000,0.562
148
+ 20,mlp.down_proj,46.70051575,0.01250,4.290
149
+ 21,self_attn.k_proj,0.00001234,0.01000,0.546
150
+ 21,self_attn.v_proj,0.00002742,0.01000,0.543
151
+ 21,self_attn.q_proj,0.00009824,0.01000,0.543
152
+ 21,self_attn.o_proj,0.00004398,0.01000,0.540
153
+ 21,mlp.up_proj,0.00063331,0.01000,0.571
154
+ 21,mlp.gate_proj,0.00064088,0.01000,0.567
155
+ 21,mlp.down_proj,202.07145691,0.01250,4.289
156
+ 22,self_attn.k_proj,0.00001307,0.01000,0.544
157
+ 22,self_attn.v_proj,0.00003479,0.01000,0.541
158
+ 22,self_attn.q_proj,0.00011022,0.01000,0.540
159
+ 22,self_attn.o_proj,0.00004273,0.01000,0.542
160
+ 22,mlp.up_proj,0.00103869,0.01000,0.571
161
+ 22,mlp.gate_proj,0.00103481,0.01000,0.563
162
+ 22,mlp.down_proj,94.00750732,0.01250,4.299
163
+ 23,self_attn.k_proj,0.00001350,0.01000,0.546
164
+ 23,self_attn.v_proj,0.00003311,0.01000,0.543
165
+ 23,self_attn.q_proj,0.00010460,0.01000,0.542
166
+ 23,self_attn.o_proj,0.00008125,0.01000,0.542
167
+ 23,mlp.up_proj,0.00153192,0.01000,0.567
168
+ 23,mlp.gate_proj,0.00151269,0.01000,0.564
169
+ 23,mlp.down_proj,37.44200134,0.01250,4.282
170
+ 24,self_attn.k_proj,0.00001306,0.01000,0.548
171
+ 24,self_attn.v_proj,0.00004170,0.01000,0.543
172
+ 24,self_attn.q_proj,0.00010361,0.01000,0.540
173
+ 24,self_attn.o_proj,0.00009263,0.01000,0.540
174
+ 24,mlp.up_proj,0.00185848,0.01000,0.567
175
+ 24,mlp.gate_proj,0.00169903,0.01000,0.562
176
+ 24,mlp.down_proj,7.99997807,0.01250,4.296
177
+ 25,self_attn.k_proj,0.00001203,0.01000,0.547
178
+ 25,self_attn.v_proj,0.00005703,0.01000,0.542
179
+ 25,self_attn.q_proj,0.00010856,0.01000,0.540
180
+ 25,self_attn.o_proj,0.00011315,0.01000,0.542
181
+ 25,mlp.up_proj,0.00322402,0.01000,0.567
182
+ 25,mlp.gate_proj,0.00282585,0.01000,0.562
183
+ 25,mlp.down_proj,1.47514915,0.01250,4.288
184
+ 26,self_attn.k_proj,0.00001247,0.01000,0.545
185
+ 26,self_attn.v_proj,0.00007589,0.01000,0.542
186
+ 26,self_attn.q_proj,0.00010745,0.01000,0.541
187
+ 26,self_attn.o_proj,0.00023876,0.01000,0.542
188
+ 26,mlp.up_proj,0.01456263,0.01000,0.568
189
+ 26,mlp.gate_proj,0.01264119,0.01000,0.564
190
+ 26,mlp.down_proj,6.14609051,0.01000,3.593
191
+ 27,self_attn.k_proj,0.00026983,0.01000,0.545
192
+ 27,self_attn.v_proj,0.00129112,0.01000,0.542
193
+ 27,self_attn.q_proj,0.00255784,0.01000,0.542
194
+ 27,self_attn.o_proj,0.03490701,0.01000,0.541
195
+ 27,mlp.up_proj,0.02056599,0.01000,0.566
196
+ 27,mlp.gate_proj,0.01958771,0.01000,0.563
197
+ 27,mlp.down_proj,0.06864744,0.01250,4.293
quantize_config.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bits": 4,
3
+ "group_size": 128,
4
+ "desc_act": true,
5
+ "sym": true,
6
+ "lm_head": false,
7
+ "quant_method": "gptq",
8
+ "checkpoint_format": "gptq",
9
+ "pack_dtype": "int32",
10
+ "meta": {
11
+ "quantizer": [
12
+ "gptqmodel:2.2.0"
13
+ ],
14
+ "uri": "https://github.com/modelcloud/gptqmodel",
15
+ "damp_percent": 0.01,
16
+ "damp_auto_increment": 0.0025,
17
+ "static_groups": false,
18
+ "true_sequential": true,
19
+ "mse": 0.0
20
+ }
21
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": "<|fim_pad|>"
25
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
3
+ size 11421896
tokenizer_config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "clean_up_tokenization_spaces": false,
199
+ "eos_token": "<|im_end|>",
200
+ "errors": "replace",
201
+ "extra_special_tokens": {},
202
+ "model_max_length": 131072,
203
+ "pad_token": "<|fim_pad|>",
204
+ "split_special_tokens": false,
205
+ "tokenizer_class": "Qwen2TokenizerFast",
206
+ "unk_token": null,
207
+ "_commit_hash": null
208
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff