Hugging Face's logo

FastFlowLM
/
Gemma3-4B-NPU2

Transformers
English
gemma3_text
unsloth
gemma3
gemma
google
Model card Files
xet
Community
FastFlowLM commited on
Commit
5567b17
·
verified ·
1 Parent(s): 93231b0

Init

Browse files
Files changed (13) hide show
  1. .gitattributes +10 -0
  2. attn.xclbin +3 -0
  3. config.json +56 -0
  4. dequant.xclbin +3 -0
  5. layer.xclbin +3 -0
  6. lm_head.xclbin +3 -0
  7. mm.xclbin +3 -0
  8. model.q4nx +3 -0
  9. tokenizer.json +3 -0
  10. tokenizer_config.json +0 -0
  11. vision_attn.xclbin +3 -0
  12. vision_mm.xclbin +3 -0
  13. vision_weight.q4nx +3 -0
.gitattributes CHANGED
@@ -33,3 +33,13 @@ 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
+ attn.xclbin filter=lfs diff=lfs merge=lfs -text
37
+ dequant.xclbin filter=lfs diff=lfs merge=lfs -text
38
+ layer.xclbin filter=lfs diff=lfs merge=lfs -text
39
+ lm_head.xclbin filter=lfs diff=lfs merge=lfs -text
40
+ mm.xclbin filter=lfs diff=lfs merge=lfs -text
41
+ model.q4nx filter=lfs diff=lfs merge=lfs -text
42
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
43
+ vision_attn.xclbin filter=lfs diff=lfs merge=lfs -text
44
+ vision_mm.xclbin filter=lfs diff=lfs merge=lfs -text
45
+ vision_weight.q4nx filter=lfs diff=lfs merge=lfs -text
attn.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad5ee7122d17040e7cbde995fe3a13fb38c77cd64921605c9b61e703fe85e070
3
+ size 464683
config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "attention_bias": false,
3
+ "attention_dropout": 0.0,
4
+ "attn_logit_softcapping": null,
5
+ "cache_implementation": "hybrid",
6
+ "final_logit_softcapping": null,
7
+ "head_dim": 256,
8
+ "hidden_activation": "gelu_pytorch_tanh",
9
+ "hidden_size": 2560,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 10240,
12
+ "max_position_embeddings": 131072,
13
+ "model_type": "gemma3_text",
14
+ "num_attention_heads": 8,
15
+ "num_hidden_layers": 34,
16
+ "num_key_value_heads": 4,
17
+ "query_pre_attn_scalar": 256,
18
+ "rms_norm_eps": 1e-06,
19
+ "rope_local_base_freq": 10000.0,
20
+ "rope_scaling": {
21
+ "factor": 8.0,
22
+ "rope_type": "linear"
23
+ },
24
+ "rope_theta": 1000000.0,
25
+ "sliding_window": 1024,
26
+ "sliding_window_pattern": 6,
27
+ "torch_dtype": "bfloat16",
28
+ "use_cache": true,
29
+ "vocab_size": 262208,
30
+ "addr_qk": 9216,
31
+ "addr_kv": 33792,
32
+ "addr_l_begin_mha": 53760,
33
+ "addr_l_end_mha": 25088,
34
+ "addr_kk": 45056,
35
+ "flm_version": "0.9.3",
36
+ "vision_model_weight": "vision_weight.q4nx",
37
+ "vision_mm_engine_xclbin_name": "vision_mm.xclbin",
38
+ "vision_mha_engine_xclbin_name":"vision_attn.xclbin",
39
+ "vision_conv2d_stride": 14,
40
+ "vision_conv2d_padding" : 0,
41
+ "vision_conv2d_kernel" : 14,
42
+ "vision_conv2d_Cin": 3,
43
+ "vision_conv2d_Cout": 1152,
44
+ "vision_average_pooling_kernel": 4,
45
+ "vision_average_pooling_stride" : 4,
46
+ "vision_average_pooling_padding" : 0,
47
+ "vision_layer_norm_eps" : 1e-06,
48
+ "vision_rms_norm_eps" : 1e-06,
49
+ "vision_intermediate_size": 4304,
50
+ "vision_hidden_size" : 1152,
51
+ "vision_head_dim": 72,
52
+ "vision_num_attention_heads": 16,
53
+ "vision_num_key_value_heads": 16,
54
+ "vision_num_hidden_layers": 27
55
+
56
+ }
dequant.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1be8893482f4d6f3634621e9f1c52355ce082cb768300ead02a54b05026b74c
3
+ size 115179
layer.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8662204bd8636a57e2ee298dff9e638e8ad257866f218cd5a7b0e3bea34f125
3
+ size 282955
lm_head.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd505faa35eb3c528ed492cb62cbc52b3f808a94896d79fe84a4bd426b209d7f
3
+ size 153355
mm.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1e88cd899e3fbef3b402ecaf2ab644f21cca2ee85dafe6e2521d89222c5a644
3
+ size 347675
model.q4nx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec52ae5c2fae20c93e40fe526fd527d630c27ecf3fa06fcf71d0e81bd0fed914
3
+ size 3768226744
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
3
+ size 33384568
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
vision_attn.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:82474ee529ed78f886566c35bdacb527667cdd25d72df628cdcda0a042db4d84
3
+ size 507643
vision_mm.xclbin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2afec6607276d44e58d845ccbd63c3170442b96871dd91675062753049dda1e7
3
+ size 186395
vision_weight.q4nx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9668714fe68024c5e7137bab0e003b6b332f4dece4229bd5dbd3d5efc350165c
3
+ size 9605883680