SorenDreano commited on
Commit
d88964b
·
verified ·
1 Parent(s): c5c6360

Upload folder using huggingface_hub

Browse files
config.json CHANGED
@@ -272,7 +272,8 @@
272
  "model.language_model.layers.30.linear_attn.in_proj_z",
273
  "model.language_model.layers.30.linear_attn.in_proj_b",
274
  "model.language_model.layers.30.linear_attn.in_proj_a",
275
- "lm_head"
 
276
  ],
277
  "kv_cache_scheme": {
278
  "actorder": null,
 
272
  "model.language_model.layers.30.linear_attn.in_proj_z",
273
  "model.language_model.layers.30.linear_attn.in_proj_b",
274
  "model.language_model.layers.30.linear_attn.in_proj_a",
275
+ "lm_head",
276
+ "re:^mtp.*"
277
  ],
278
  "kv_cache_scheme": {
279
  "actorder": null,
model.safetensors.index.json CHANGED
@@ -1,7 +1,6 @@
1
  {
2
  "metadata": {
3
- "total_parameters": 4539265808,
4
- "total_size": 6520006176
5
  },
6
  "weight_map": {
7
  "model.language_model.embed_tokens.weight": "model-00001-of-00002.safetensors",
@@ -998,6 +997,21 @@
998
  "model.visual.merger.norm.weight": "model-00002-of-00002.safetensors",
999
  "model.visual.patch_embed.proj.bias": "model-00002-of-00002.safetensors",
1000
  "model.visual.patch_embed.proj.weight": "model-00002-of-00002.safetensors",
1001
- "model.visual.pos_embed.weight": "model-00002-of-00002.safetensors"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1002
  }
1003
- }
 
1
  {
2
  "metadata": {
3
+ "total_size": 6761330736
 
4
  },
5
  "weight_map": {
6
  "model.language_model.embed_tokens.weight": "model-00001-of-00002.safetensors",
 
997
  "model.visual.merger.norm.weight": "model-00002-of-00002.safetensors",
998
  "model.visual.patch_embed.proj.bias": "model-00002-of-00002.safetensors",
999
  "model.visual.patch_embed.proj.weight": "model-00002-of-00002.safetensors",
1000
+ "model.visual.pos_embed.weight": "model-00002-of-00002.safetensors",
1001
+ "mtp.fc.weight": "model_mtp.safetensors",
1002
+ "mtp.layers.0.input_layernorm.weight": "model_mtp.safetensors",
1003
+ "mtp.layers.0.mlp.down_proj.weight": "model_mtp.safetensors",
1004
+ "mtp.layers.0.mlp.gate_proj.weight": "model_mtp.safetensors",
1005
+ "mtp.layers.0.mlp.up_proj.weight": "model_mtp.safetensors",
1006
+ "mtp.layers.0.post_attention_layernorm.weight": "model_mtp.safetensors",
1007
+ "mtp.layers.0.self_attn.k_norm.weight": "model_mtp.safetensors",
1008
+ "mtp.layers.0.self_attn.k_proj.weight": "model_mtp.safetensors",
1009
+ "mtp.layers.0.self_attn.o_proj.weight": "model_mtp.safetensors",
1010
+ "mtp.layers.0.self_attn.q_norm.weight": "model_mtp.safetensors",
1011
+ "mtp.layers.0.self_attn.q_proj.weight": "model_mtp.safetensors",
1012
+ "mtp.layers.0.self_attn.v_proj.weight": "model_mtp.safetensors",
1013
+ "mtp.norm.weight": "model_mtp.safetensors",
1014
+ "mtp.pre_fc_norm_embedding.weight": "model_mtp.safetensors",
1015
+ "mtp.pre_fc_norm_hidden.weight": "model_mtp.safetensors"
1016
  }
1017
+ }
model_mtp.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f993d7b896c6d3c72ee66fd446b28bcf316d5f5ce4a0427c0442dfe461cbe1b
3
+ size 241200704